University of Southern California Dissertations and Theses

A Study Of The Comparability Of The Wisc And The Wais And The Factors Contributing To Their Differences

(USC Thesis Other)

A Study Of The Comparability Of The Wisc And The Wais And The Factors Contributing To Their Differences

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content This dissertation has been microfilmed exactly as received 68-17,043 SIMPSON, Robert Lee, 1921- A STUDY OF THE COMPARABILITY OF THE WISC AND THE WAIS AND THE FACTORS CONTRIBUTING TO THEIR DIFFERENCES. University of Southern California, Ph.D., 1968 Education, psychology University Microfilms, Inc., Ann Arbor, Michigan A STUDY OF THE COMPARABILITY OF THE WISC AND THE WAIS AND THE FACTORS CONTRIBUTING TO THEIR DIFFERENCES by Robert Lee Simpson A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (Education) June 1968 UNIVERSITY O F SOUTHERN CALIFORNIA THE GRADUATE SCHOOL UNIVERSITY PARK LOS ANGELES. CALIFORNIA 9 0 0 0 7 This dissertation, w ritten by under the direction of h..is.. Dissertation C om mittee, and approved by all its members, has been presented to and accepted by the Graduate School, in partial fulfillm ent of requirements for the degree of D O C T O R O F P H I L O S O P H Y Dean Date. DISSERTATION CQMMITTEE Chairman ACKNOWLEDGEMENTS I wish to express my appreciation to Professor Donald Ralph Schrader, my major professor, for his guidance and counsel in my doctoral program. The assistance of my committee, Dr. Calvin C. Nelson, Dr. Constance Lovell, Dr. C. Edward Meyers, Dr. Newton S. Metfessel and Dr. D. Welty Lefever, was greatly appreciated. Special mention is due to Dr. Ralph Hoepfner who gave generously of his time and made valuable suggestions on several occasions. I wish to express my gratitude to Angus MacLean for his assistance with the statistical procedures, to John Dennis for his work with the computer and to Mrs. Roberta Riley for her typing skills. I am very grateful to my wife, Marjorie, for her en couragement, patience, and understanding during the six years that the doctoral work has required. My five children, Paul, Richard, Roger, Barbara, and James have perhaps paid the great est price in having a part-time father during this time. TABLE OF CONTENTS Page ACKNOWLEDGEMENTS........................................... ii LIST OF TABLES........................................... vi LIST OF FIGURES.......................................... viii Chapter I. THE PROBLEM AND DEFINITIONS OF TERMS USED........... 1 Introduction The Problem Statement of the Problem Importance of the Study Definitions of Terms Used I.Q. Verbal I.Q. Performance I.Q. Full Scale I.Q. Dull Normal EMR TMR Organization II. REVIEW OF THE LITERATURE......................... 12 History of Mental Testing Classification of Mental Retardates Academic, Social and Vocational Success of Retardates Regular Versus Special Class Placement for EMR's Cultural Differences in Intelligence Racial Differences in Intelligence Description of the WISC and the WAIS Standardization of the WISC and the WAIS WISC Versus WAIS Reliability WISC-WAIS Validity Studies Subtests of the WISC and the WAIS Information Comprehension Arithmetic Similarities iii / TABLE OF CONTENTS (continued) Page Block Design Object Assembly Digit Symbol (Coding). Sub'test Contribution to I.Q. Diagnostic and Clinical Features Summary of the Review'of the Literature III. METHOD......................................... 73 Selection of the Sample Source Population Number OneJ Purpose Description Selection of the Sample The Testing of Sample Number One Test Examiner Place of Testing Testing Procedure The Data from Sample One Preparation of the Data for Analysis Testing Hypothesis Number One The Practical Effects on Placement Recommendations of WISC-WAIS Test Differences Subtest Contributions to WAIS-WISC Differences Population Number Two Description of Population Number Two The Sample from Population Number Two Gathering the Data Preparation of the Data Testing Hypothesis Number Two IV. FINDINGS ..................................... 94 Hypothesis One Verbal I.Q.’s for Sample One Performance I.Q.'s for Sample One Full Scale I.Q.'s for Sample One Summary of Significant Analysis of Variance Main Effect WAIS Minus WISC Differences Product-Moment Correlations Between WISC and WAIS WISC-WAIS Differences in Placement Recommendation WISC Mean Subtest Scale Scores WAIS Mean Subtest Scale Scores Subtest Contributions to WISC-WAIS I.Q. Differences Subtest Contribution Differences by Race iv TABLE OF CONTENTS (continued) Page Comparison of the I.Q.'s of Sample One and Sample Two Hypothesis Number Two I.Q. and Reading Index Correlations with Graduate Status Test of Significance of Differences between Correlations I.Q. Versus Reading Index as Predictors of Success Graduate Status in Regular Classes by I.Q. Ranges Graduate Status in EMR Classes by I,.Q. Ranges V, DISCUSSION, CONCLUSIONS, AND RECOMMENDATIONS....... 125 Discussion Test of Hypothesis One Entire Sample Race Groups Correlations Effect on Placement Subtest Contribution Test of Hypothesis Two Conclusions Recommendations VI. SUMMARY....................................... 138 APPENDIX A.............. 148 APPENDIX B............................................. 155 APPENDIX C............................................. 159 APPENDIX D............................................. 167 BIBLIOGRAPHY. ......................................... 170 v LIST OF TABLES Table Page 1 Distribution of Standardization Sample in Composite Stanford-Binet I.Q. on Forms L and M.............. 15 2 Expectancies at Various Levels of Mental Abilities. . . 16 3 Educational Records of 2500 Seventh Graders ......... 21 4 ' Compton Union High School District Distribution of Attendance ........................... 75 5 Verbal I.Q.'s for Sample One........................ 94 6 Performance I.Q.'s for Sample One.................. 95 7 Full Scale I.Q.'s for Sample One.................... 97 8 Summary of Significant Analysis of Variance Main Effects arid Interactions Pertaining to the Comparability of the WISC and WAIS................ 98 9 WAIS Minus WISC Differences....................... 99 10 Product-Moment Correlations Between WISC and WAIS for Sample One.............. . . 100 11 WISC-WAIS Differences in Placement Recommendation. . . . 102 12 | WISC Mean Subtest Scale Scores............ 104 13 WAIS Mean Subtest Scale Scores........................106 14 Subtest Contributions to WAIS-MINUS-WISC I.Q. Differences for All Subjects ..................... 108 15 Subtest Contributions to WAIS-MINUS-WISC I.Q. Differences by Race........ 109 16 Reading Achievement and I.Q.'s of Sample Two........... Ill 17 Comparison of the I.Q.'s of Sample One and Sample Two........................................112 18 Comparison of the Regular and EMR Class Students in Sample Two................. 113 vi LIST OF TABLES (Continued) Table Page 19 Comparison of Graduates and Dropouts in Regular Class Placement............................ 114 20 I.Q. and Reading Index Correlations with Graduate Status .................................. 115 21 Significance of the Differences Between I.Q. and Reading Index Correlations with Graduate Status for Regular Class Students.......................... 116 22 Significance of the Differences between I.Q. and Reading Index Correlations with Graduate Status for EMR Students................................... 117 23 I.Q. Versus Reading Index as Predictors of Success in Regular Classes ................................118 24 Combined WISC and WAIS Full Scale Versus Reading Index. . 118 25 Students Graduating in Regular Classes by I.Q. Ranges . . 120 26 Students Graduating in EMR Classes by I.Q. Ranges .... 120 27 Students Graduating in Regular Classes by Reading Index Ranges........................... ..122 28 Students Graduating in EMR Classes by Reading Index Ranges................................123 vii LIST OF FIGURES Figure Page 1 Comparison of I.Q.'s of Regular and EMR Class Graduates. 121 2 Comparison of Reading Indices of Regular and EMR Class Graduates..............................124 viii CHAPTER I THE PROBLEM AND DEFINITION OF TERMS Introduction The field of education is being greatly influenced by rapid tech nological changes taking place today. Drastic reduction in the un skilled and semi-skilled labor markets, the population explosion, and the higher age of compulsory school attendance requirements have changed the character of the present high school population compared with that of a generation ago. Previously the ill-equipped and poorly motivated, student who dropped out of school had some reasonable chance of em ployment^ but this is no longer the case. Educators are struggling with the complex teaching-learning problems produced by large numbers of low- achieving students in public and private schools. The utopian educational process would involve one teacher and one student, but economic considerations and teacher availability require that each teacher instruct a number of students at a time. Teaching a number of students simultaneously produces problems; because students do not all learn at the same rate or in the same manner, nor do all stu dents profit equally from a given medium. The teacher of many cannot gear his presentations to meet individual needs as the teacher of one could do. In most grades in most schools there are students who fall farther and farther behind the others. For some it is possible that the meager benefits in terms of learning that they obtain from school 2 attendance may be outweighed by the psychological damage of continued frustration and failure. It is believed that these students can be subclassified into two types: (1) those who could achieve and develop academically if certain learning deficits were removed and if they could be peimitted to proceed more slowly and with more help on the basic skills, and (2) those who, because of limited intellectual endowment, could not be expected to master the academic material in regular classes no matter how hard they tried and no matter how much extra help was given them. Students of the first type are called ’ ’ slow learners" and the second "mentally re tarded." Kough and DeHaan (1955, p. 70) have listed some indentifying characteristics of these below-average ability students: 1. unable to think abstractly or to handle symbolic material 2. unable to understand and carry through directions for assignments 3. lacks the "common sense" and reasoning levels of the group 4. unable to understand complex game rules 5. slow in all areas: academic, social, emotional, and physical 6. breaks rules of conduct or of games and is often unaware of it 7. unable to work independently 8. easily confused 9. short interest and attention span 10. unable voluntarily to concentrate 11. finds it extremely difficult, if not impossible, to keep up with the class on academic work 12. behind noxmal grade achievement in school Distinguishing between mentally retarded and slow learning chil dren is often difficult and requires a thorough examination of the child in teims of several areas of development. Furtheimore, the diag nosis of children of either type has far-reaching implications for their future educational experiences. Mentally retarded students in regular 3 classes may suffer loss of self-esteem as a result of academic chal lenges which are beyond their capabilities. On the other hand, slow learners, who may have near-average potential, may resign themselves to the performing level of their mentally retarded classmates when they are placed in special classes. Responsibility for determining whether children are mentally retarded or slow learners has been given to indi viduals with specialized training in this area and certificated by the State of California as school psychologists. The California Legislature has in very general terms described the characteristics of the mentally retarded for special class place ment but has looked to the field of psychology to assume the task of operationally defining and further delineating their characteristics. Mentally retarded minors means all minors who because of retarded intellectual development as determined by an individual psychological examination are incapable of being educated efficiently and profitably through ordinary classroom instruction. (Education Code Section 6901) In 1965 Campbell, Daly and Hanson presented a Bulletin under the auspices of the California State Department of Education which read: The California law relating to programs for the educable mentally retarded minors, as set forth in Education Code 6908, places the major respon sibility for identifying these minors upon the school psychologist. Therefore, it is most im portant that all school systems charged with the responsibility of establishing and maintaining special programs for educable mentally retarded minors obtain the services of a well-qualified school psychologist, who is trained in modern clinical methods, understands the problems of mental retardation, and appreciates the prob lems of school organization and classroom 4 instruction. The school psychologist must meet certification requirements set forth by the State Board of Education. (Campbell, Daly, and Hanson, 1965) Heber wrote a manual on mental deficiency for the American Asso ciation of Mental Deficiency in which the operational definition of mental retardation was "subaverage general intellectual functioning which originates during the developmental period (conception to age 16) and is associated with impairment in adaptive behavior." "Subaverage" indicated performance which was more than one standard deviation below the average as measured by an intelligence test. This would mean below I.Q. 84 on the 1960 revision of the Stanford-Binet and below 85 on the Wechsler Scales. "Impairment in adaptive behavior" referred to the effectiveness of the individual in adapting to the natural and social demands of his environment. Heber went on to say that impaired bfeha- vior may be reflected in (1) maturation, (2) learning, and/or (3) social adjustment. These three aspects of adaptation are of differ ent importance as qualifying conditions . . . for different age groups . . . rate of maturation re fers to the rate of sequential development of self-help skills in infancy and early childhood, such as sitting, crawling, standing, walking, talking, habit training, and interaction with group age peers. In the first few years of life adaptive behavior is assessed almost completely in terms of these and other manifestations of sensory-motor development. Consequently, delay in acquisition of early developmental skills is of prime importance as a criterion of mental re tardation during the preschool years. "Learning ability" refers to the facility with which know ledge is acquired as a function of experience. Learning difficulties are usually most manifest in the academic situation and if mild in degree, 5 may not even become apparent until the child enters school. Impaired learning ability is, therefore, par ticularly important as a ! qualifying condition of mental retardation during the school years. "Social adjust ment" is particularly important as a qualifying condi tion of mental retardation at the adult level where it is assessed in terms of the degree to which the indi vidual is able to maintain himself independently in the community and in gainful employment as well as by his ability to meet and conform to other personal and social responsibilities and standards set by the community. During preschool and school age years social adjustment is reflected, in large measure, in the level and manner in which the child relates to parents, other adults, and age peers. (Heber, 1959) Campbell (1965) suggests that school psychologists, for purposes of administrative guidelines in recommending students for educable mentally retarded placement, use the I. Q. range of 50 to 79 and that the social and physical developmental history be given equal weight with that of intellectual functioning as derived from verbal and non verbal intelligence test scores. In spite of these recommendations, there persists a widespread tendency on the part of school psychologists to base placement in EMR classes primarily, and at times almost exclusively, on I. Q. (Webb, 1964). This may be due to the fact that there are several standardized intelligence tests which provide objective measurement., .whereas similar instruments measuring social effectiveness and relative physical devel opment in numerical terms are not as easily standardized. The Problem Statement of the Problem The two major objectives of this research, stated in general terms, were to contribute to a greater understanding: (1) of the 6 comparability of the Wechsler Intelligence Scale for Children (WISC) and the Wechsler Adult Intelligence Scale (WAIS), and (2) of the rela tive predictive accuracy of these tests as compared to that of individual reading achievement tests. More specifically, the purpose of the study was to seek answers to the following questions: 1. Are there significant differences between the WISC and WAIS Verbal, Performance, and Full Scale I. Q.’s of below-average intel ligence Anglo, Mexican-American, and Negro-American boys and girls? 2. What is the practical effect on class placement of any score differences between the WISC and the WAIS? 3. Which subtests account for the differences in scores between the two tests? 4. How do the results of individual intelligence tests compare to those of individual reading achievement tests as predictors of suc cess in regular high school classes for suspected mental retardates? The independent variables investigated in the study were test differences, race, and sex. In Part 1 the dependent variables were the Verbal, Performance, and Full Scale I. Q.'s and the subtest scale scores for each race and sex group. In Part 2 the independent variables were I. Q. and reading index, the dependent variable was outcome with regard to success or failure in high school. To study the relationships between the independent and dependent variables two major hypotheses were established and are presented here in order by which they are discussed in later chapters: 7 1. WAIS will exceed WISC I.Q.'s for the sample of below average intelligence subjects as a whole, for all race, sex, and race-sex groups on the Verbal Scale, Performance Scale and Full Scale. 2. Individual reading achievement tests will predict more accurately than individual I.Q.'s with refer ence to success in graduating from high school in regular classes by suspected mental retardates. Importance of the Study For the past five years the investigator has been employed by the Compton Union High School District as a school psychologist. The three psychologists employed by the District have two major responsibilities which require the greatest share of their time. The first responsi bility is that of testing students who have been unsuccessful in the regular class program and who have been referred for evaluation to de termine whether they qualify for placement in special training classes for the educable mentally retarded (EMR). The second major responsi bility is retesting students who are already in EMR classes and who have not been re-evaluated within the preceding three year period. In both types of testing situation the student is administered a battery of tests which includes the WISC, if the subject is less than sixteen years of age, or the WAIS, if he has passed his sixteenth birthday. Although other factors are considered, students scoring 80 or above on the Full Scale of the intelligence test are usually recommended for con tinued regular class placement, while those scoring below 80 are usually recommended for placement in educable mentally retarded classes. Some exceptions to this practice occur when a below-eighty Full Scale I.Q. is produced by a relatively high score on one of the subscales (Verbal I.Q. 8 or Performance I. Q.) and a relatively low score on the other. Students scoring below 55 and who are judged to be unable to profit from the EMR program may be placed in special classes for the trainable mentally re tarded (TMR). Another kind of exceptional case is when a low level of performance on the intelligence test is attributable primarily to a vision or hearing handicap or to bilingualism. It should be kept in mind that only students who have been unable to do regular classwork adequately are referred for testing. There are undoubtedly some students who would score in the EMR range if they were tested, but they have ma naged to get passing grades in regular classes. Teachers vary greatly in their grading practice even when the school administration provides fairly clear-cut guidelines. It is probable that well-behaved, con forming mentally retarded youngsters who appear to be doing as well as they can are much less likely to be referred for EMR evaluation than are mental retardates who misbehave. Many psychologists using the two Wechsler intelligence tests at the high school level have been disturbed by the frequency with which students have been placed in EMR classes on the basis of Stanford-Binet or WISC scores in the upper elementary grades or in junior high school only to be found ineligible for continued EMR placement because of an I. Q. well above the EMR range (79) when retested several years later on the WAIS. Possible explanations which have been offered for this phenomenon include (1) regression effects (Webb, 1964): the tendency for subjects getting high or low scores on a first test administration to get scores closer to the mean on a second administration, 9 (2) testing effects: the tendency to do better on a second administra tion of a test than on the first administration as a result of learning talcing place from the experience of talcing the test, (3) that placement in special classes for several years increases the subject's ability to do well on the test, (4) that one (or both) of the tests is unreliable, and (5) that the two tests are not equal in teims of difficulty, at least for the below average student. An examination of one section of the District clinical case files produced 42 cases of students who had been in EMR classes for a period of three years or more and who had been returned to regular classes af ter obtaining a retest I. Q. well above the EMR range. We were inter ested in. knowing what had happened to these students who had previously been unable to cope with regular classes and who had experienced sev eral years of less-demanding EMR classes but now had to return to regular classes. How many could survive regular classes long enough to graduate from 'grade twelve? In seventeen of the forty-two cases the answer was unavailable because the students had transferred out of the district or were not old enough to have graduated. However, of the remaining twenty-five cases, twenty-one (84 per cent) had dropped out of school, while only four (16 per cent) had graduated. It is important that psychologists and administrators obtain a more accurate under standing of the comparability of the I. Q. tests used in placement decisions at the high school level. 10 Definition of Terms Used I. Q.--Intelligence Quotient. " . . . the ratio between a parti cular score which an individual attains (on a . given intelligence test) and the score which an average individual of his life age may be assumed to attain on the same test, when both scores are expressed in the same notation." (Wechsler, 1958, 28) Verbal I. Q.--Intelligence quotient provided by the combined con tributions of the Information, Comprehension, Arithmetic, Similarities, Digit Span, and Vocabulary subtests of the WISC or the WAIS. Perfoimance I. Q.--Intelligence quotient provided by the com bined contributions of the Picture Completion, Picture Arrangement, Block Design, Object Assembly, and Coding (or Digit Symbol in the case of the WAIS). Full Scale I. Q.--Intelligence quotient provided by the combined contributions of all of the Verbal and Perfoimance subtests of the WISC or the WAIS. It is the Full Scale I. Q. which has greatest influence in determining class placement. Dull Normal.--A range of 10 I. Q. points between 80 and 89. In dividuals with I. Q.'s in this range are not low enough to qualify for EMR class placement but generally are not able to do average work in school in the regular classes. EMR.--Educable mentally retarded. In terms of I. Q., a range of scores between 50 and 79. In terms of academic aptitude, an individual who cannot be expected to keep up with even slow-moving regular classes, but who can be taught to read, write, and calculate well enough for most tasks involved in everyday living. 11 TMR. - -Trainable mentally retarded. Having an I. Q. below 50. Unable to profit much at all from academic training. Unable to learn to read meaningfully beyond a few words. Can be trained to at least par tially support oneself in some menial kind of work. Reading Index.--A ratio which expresses one's level of reading achievement in relation to actual grade placement. Obtained by dividing reading achievement by actual grade placement. Organization Five additional chapters are included in this study. Chapter II reviews the related literature. Chapter III describes the experimental design, materials, methods, and procedures. Chapter IV presents the findings of the study. Chapter V is devoted to discussion, conclusions, and recommendations. Chapter VI summarizes the study. CHAPTER II REVIEW OF THE LITERATURE History of Mental Testing Darwin (1859) proposed in his theory of natural selection that the . various species of animals survived and reproduced their kind only as their inherited characteristics enabled them to cope with their en vironment. The extinction of some species may be attributable to their failure to adjust to a changed environment. This ability to cope with environment has been suggested as one definition of intelligence. Sir Francis Gal ton was one of the early students of individual differences in intelligence. He believed that intelligence was in herited (1869). Cattell studied individual differences in the laboratory and be lieved that he could identify superior intelligence by means of tests of sensory acuity, grip strength, and memory for dictation. Spearman (1904) maintained that intellectual activities have in common a single general "g" factor. Each test was assumed to measure some "g", some specific "s" factor, and some error "e" factor. Probably the greatest single contribution in the field of mental testing was made by Binet and Simon who were commissioned by the Paris Minister of Education to devise a way of distinguishing between students who could profit from regular academic instruction in the schools from 12 13 those who could not. In response to this commission the first Simon- Binet Scale was introduced in 1905, and revisions were published in 1908 and 1911. Binet believed that nearly all of the phenomena with which psycho logy concerned itself were phenomena of intelligence, but he considered it useless to examine such things as sensations and motor responses in intelligence tests (Binet and Simon, 1916). Intelligence was considered to be a fundamental faculty, the alteration of which was of the utmost importance for practical life. This faculty was judgment, otherwise called good sense, initiative, the faculty of adapting oneself to cir cumstances. The ability to judge well, to comprehend well, to reason well; these were the essential activities of intelligence for Binet. To him, the rest of the intellectual faculties were of little importance in comparison with judgment. (Hunt, 1961, 12) In the United States the great leap forward in mental testing oc curred in 1916 when Louis M. Terman of Stanford University, after six years of experimentation, produced the StanforcL-Binet revision of the French test. Whereas Simon and Binet had emphasized the measurement of intelligence in suspected mental defectives, Teiman's version included provision for assessing the intellectual abilities of normal and super ior children as well. (Cronbach, 1960, 181) The Stanford-Binet achieved almost immediate and widespread popularity and has been improved with revisions in 1937 and 1960. It has been translated into many lan guages and has served as the "criterion” for the development and the validity studies of most other intelligence tests to be developed since 14 1916. Some of the important features of the Stanford-Binet are: 01) It measures present ability rather than inborn capacity. (2) It is very useful for predicting certain kinds of behavior. (3) It is strongly weighted with verbal activities. (4) It measures somewhat different mental abilities at different ages. (5) It requires experiences common to the urban culture of the United States and is of dubious value for comparing cultural groups. (6) It does not give a reliable measure of separate aspects of mentality. (7) Its score is influenced by the sub ject's personality and emotional habits (Cronbach, 1960, 181). Terman (1916) viewed intelligence in terms of the ability to achieve in school. He found that children with I. Q.'s below 85 were typically retarded in school by one or two years and that success in school and level of in telligence were highly related to success in most adult areas (Terman et al.3 1925). Terman's use of the concept of intelligence quotient represented a major factor in standardizing the testing of general men tal ability. Table 1 shows the percent of individuals found in the various I. Q. ranges in the Stanford-Binet standardization sample. It will be noted that slightly less than half (46.5 per cent)of this population falls within the average range (90-109). In the normal curve it would be expected that half of the remaining members of the population would be above the average range and half would be below. According to Table 1, however, 30.73 per cent was above the average range and only 22.73 per cent was below. 15 TABLE 1 DISTRIBUTION OF STANDARDIZATION SAMPLE IN COMPOSITE STANFORD-BINET I. Q. ON FORMS L AND M I. Q. Percentage of Cases Classification 160-169 0.03 150-159 0.2 Very superior 140-149 1.1 130-139 3.1 Superior 120-129 8.2 110-119 18.1 High average 100-109 23.5 Normal or average 90-99 23.0 80-89 14.5 Low average 70-79 5.6 Borderline defective 60-69 2.0 50-59 0.4 Mentally defective ' 40-49 0.2 30-39 0.03 Merrill, 1938, 650; and Anastasi, 1959, 198. Cronbach has compiled the data from a number of investigations of the relationship between intelligence and educational and vocational potential. Table 2 shows the expectancies for individuals of various I. Q. ranges. The present trend toward requiring students to remain in school to higher age levels coupled with the increasing efforts to adapt the curriculum to individual students'needs and abilities will probably render some of these figures obsolete. For example, Table 2 indicates that a student with an I. Q. of 75 has only about a 50-50 chance of reaching high school. This is probably no longer true today and almost certainly will not hold in the future. TABLE 2 EXPECTANCIES AT .VARIOUS LEVELS OF MENTAL ABILITIES I. Q......................................... ........... ....... 130 Mean of persons receiving Ph. D. 120 Mean of College graduates 115 Mean of Freshmen in typical four year college Mean of children from white collar and skilled labor homes 110 Mean of high school graduates Has about 50-50 chance of graduating from college 105 About 50-50 chance of passing in academic high school curriculum 100 Average for total population 90 Mean of children from low-income city homes or rural homes Adult can perform jobs requiring some judgment (operate sewing machine) 75 About 50-50 chance of reaching high school Adult can keep small store, perform in orchestra 60 Adult can repair furniture, harvest vegetables, assist electrician 50 Adult can do simple carpentry, domestic work 40 Adult can mow lawns, do simple laundry (Cronbach, 1960, 174) Mental tests have been found useful for prediction and classifi cation in a number of. areas such as the public schools, industry, and the armed forces. A severe limiting factor in the use of such tests as the Stanford-Binet, particularly where large numbers of people are involved, has been the amount of time required for the individual ad ministration. This has led to the development of group tests which can be administered to a relatively large group of subjects at one time and generally do not require as highly trained examiners as the individual tests do. The greatest expansion of group testing occurred in World War I when the United States Army tested one and one-half million men. 17 The Army General Classification Test (AGCT) was used for this purpose. Since this test required the ability to read on the part of the subjects, it was found inadequate for the testing of illiterates, and a second (Beta) form was devised for assessing the intelligence of non-readers. Although the Stanford-Binet Scales provided a reasonably good assessment of verbal abilities, it was less useful with subjects who were mentally retarded, emotionally maladjusted, or psychotic. This limitation led to the development of a number of performance scales, such as the Healy Completion, the Pintner-Paterson series, and the Porteus Mazes. The most significant tests of performance were developed by David Wechsler, a clinical psychologist on the staff of Bellevue Hospital in New York City. He introduced the Wechsler-Bellevue Form I in 1939 to provide for clinical evaluation of social derelicts, feeble minded, psychotic, and illiterate. As in the case of the Stanford- Binet, the Wechsler scale found quick acceptance and was of great value in military hospitals in World War II. In 1946 a second form of the test was published but was never adequately standardized (Croribach, 1960, 192). Probably part of the reason for the failure to standardize the second form of the scales more completely was that it was being re placed by a new version which would be usable with younger children. The development and standardization of this new test, the Wechsler In telligence Scale for Children (WISC) as well as the greatly improved revision and modification of the Wechsler Form I which is now known as the Wechsler Adult Intelligence Scale (WAIS), will be discussed in a later section. 18 Classification of Mental Retardates As Anastasi (1958, 378) has pointed out, the concept of mental de ficiency is closely related to the definition of intelligence. She defines mental deficiency as a pronounced inferiority in those abilities that are essential for survival in our cultural milieu. Therefore, suc cessful accomplishment in school, ability to support oneself economically and performance on the standardized I. Q. tests have been especially prominent in our understanding and definition of mental deficiency. There have been various criteria suggested for the classification of mental retardates, but in general they tend to fall into four general categories: (1) Psychometric definitions are based on the individual’s per formance on an intelligence test. One of the earliest of this type was suggested by Terman, who based it on the data obtained in standardizing the first Stanford-Binet Scale. His five categories of retardates, along with their I. Q. ranges, were: Dullness 80-90 Borderline 70-80 Moron 50-70 Imbecile 20-50 Idiot below 20 (Terman, 1916) These specified ranges have achieved wide acceptance, but unfor tunately, as Anastasi (1958, 380) has pointed out, the original Stanford- Binet Scale had a standard deviation of only twelve I. Q. points, while the 1960 revision has a standard deviation of sixteen, the Wechsler tests fifteen, and other widely known tests such as the California Test of Mental Maturity (a group test) a standard deviation of twenty. This 19 leads to some confusion in interpreting the various tests. (2) The British have urged an approach to classification of men tal retardation based on legal criteria. Tredgold and Soddy (1956, 5) define mental retardation as "a condition of arrested or incomplete de velopment of mind existing before the age of eighteen years, whether arising from inherent causes or induced by disease or injury.” They describe the moron as one who requires care, supervision, or control for his own protection or for the protection of others; the imbecile as one having a mental defect which prevents him from managing his own affairs; and the idiot as one who is unable to guard himself against common physical dangers. (3) Doll (1953) used the social competence approach and developed the Vineland Social Maturity Scale which produces a social age and a social quotient using an interview technique. Obtaining data by this method involves collecting information either from the subject or an informant regarding the subject's ability to perform everyday living tasks divided into categories: general self-help, self-help in eating, self-help in dressing, self-direction, occupation, communication, loco motion, and socialization. Anastasi (1958) has pointed out that the Vineland and the Stanford-Binet correlate sufficiently low to indicate that they are measuring differing aspects of behavior. The Vineland is particularly useful in decisions regarding institutionalization, but not in decisions regarding a choice of educable mentally retarded or train- able mentally retarded class placement. (4) In recent years there has been an increasing tendency on the 20 part of the public schools to take on the responsibility of developing the abilities of retardates; whereas formerly the retardate dropped out of school after he had fallen hopelessly behind. Educational classifi cation of retardation, while relying to a considerable extent on the performance on an intelligence test and on the demonstrated degree of social effectiveness in the school environment, also uses as a criter ion the child's ability to profit from various specified levels of in struction as found in regular classes, low-regular classes, special classes for the educable mentally retarded, and special classes for the trainable mentally retarded. Thus for school placement, typical cate gories and criteria are as follows: Category Slow learners I.Q. Range 80-89 Educable mentally retarded 50-79 Trainable mentally retarded. 25-49 Description of Abilities Unable to keep up with regular classes when the material is pre sented at the usual rate. Placed in slower-moving classes with em phasis on basic skills and concrete experiences. High School gradua tion is possible but not likely in most cases. Not college potential. Unable to profit much at all from regular classes and little from low regular classes even with con siderable help. In these (EMR) classes the emphasis is on devel oping ability to manage own home, get and hold a job (unskilled or semi-skilled). High School gradu ation is very remote except in special-training classes where group standards are not imposed. Unable to learn to read or write even with extensive individual tu toring, but may be able to recog nize crucial words such as 21 "danger," "detour." Emphasis on learning to get from place to place, staying out of trouble, sup porting oneself, at least partially, with menial job, looking after one's own health, hygiene, and safety. Severely mentally below 25 Unable to care for self or to pro retarded fit from any kind of instruction. Requires constant supervision in home or institution. No provision for this type in public schools. Academic, Social, and Vocational Success of Retardates In 1949 Dillon published the results of a survey of five school systems with reference to the problem of early school dropouts. One of the most impressive features of his presentation was the following table which related dropout rate to intelligence in the junior and senior high school years: TABLE 3 Education Records of 2500 Seventh Graders Intelligence Quotient Below 85 85-94 95-104 105-114 115+ All cases in grade 7 400 575 650 575 400 Dropouts in grades 7 § 8 93 30 14 5 2 Remainder entering grade 9 307 545 636 570 398 Dropouts in grades 9 § 10 241 171 143 78 29 Remaining in grade 11 66 374 493 492 369 Dropouts in grades 11 § 12 52 65 81 55 25 Remainder graduating 14 309 412 437 344 (% of original cases graduating)*3.5 53.7 63.4 76.0 86.0 Source: Dillon, 1949 ^Percentages added by present investigator 22 Cronbach (1949, 172), referring to the above table, pointed out that the very dull tend to drop out as soon as they reached age 16; and this usually meant, since they were usually retarded one or two grades, that they left before the ninth grade. The table indicated that, for this sample at least, almost no one having an I.Q. below 85 remained to graduate at the end of the twelfth .grade. Only two years before Dillon's study was published the California Legislature made it mandatory for county superintendents of schools, elementary school districts, and unified districts to provide for the education of educable mentally retarded pupils in special training classes (California Education Code Sections 6902, 6904, and 8951). In 1949 the California Legislature expanded provision for educable mental retardates to high school age students on a permissive basis (California Education Code Section 6905). In 1956, it was made mandatory for high school districts with more than 900 average daily attendance to maintain such classes for the educable mentally retarded (California Education Code Section 6904). In 1961, county superintendents were given per mission to operate classes at the high school level for districts having less than 900 average daily attendance (California Education Code Section 8951b). Dunn (1963, 110), has pointed out the disadvantages of keeping retardates in the same classes with average and above-average students. Summarizing the research on the subject of regular versus special class placement he found that retardates who remain in the regular grades slip further and further behind until finally their gains in 23 achievement are only about one third of expectancy. Meanwhile, the brighter students have been neglected and the mental health of the teacher and the retarded pupils badly strained. Dunn believes that special class placement for the educable mentally retarded can be jus tified on four major counts: (1) to provide a more adequate curriculum which stresses the development of social and vocational skills as well as the acquisition of skills in the basic tool subjects, (2) to enable more individual instruction because of the smaller classes for EMR's, (3) to remove the pressure on the retarded through reducing failure, and (4) to enable the regular class teacher to give more attention to average and bright students when the range of individual differences is reduced. Sarason (1955, 555) has pointed out that the criteria ordinarily used to identify mental retardates are not equally effective in pre dicting social and vocational success except at the extremes. It is reasonably safe to predict that a borderline retardate will never achieve the higher professional levels, and it is equally safe to pre dict that a severely retarded individual will never be able to live in complete independence in society, but between these extremes, accurate prediction becomes very doubtful. In the Onondaga County Survey (Sarason, 1955, 555) it was found that the prevalence of retardation increased steadily until the age where compulsory school attendance stopped, and at that point there was a sharp decrease in retardation. Sarason believes that most individuals who are identified as retarded in the school environment--perhaps as many as two out of three--are able 24 to make a satisfactory adjustment once out of school. Surveying the literature on mental retardates, Shepard (1966) found that mentally retarded adults usually become occupationally placed under ordinary circumstances, but change jobs frequently during the early employment years, and are usually the first to lose their jobs. They often change jobs because of personal preference rather than for other reasons, and they hold jobs that are classified as unskilled or semi-skilled. Sheperd suggested that the adjustment of mentally retar ded adults is highly dependent upon personal characteristics, the economic situation of a community, and the community attitudes toward the retardate. The consensus of studies by Bobroff (1956), Cassidy and Phelps (1955), Collman and Newlyn (1957), Flescher (1956), McKeon (1946), and Peckham (1951) was that vocational failure of retardates was usually due to personality problems. Goldstein (1957) summarized studies by Bailer (1936), Charles (1953), Fairbanks (1933), and Kennedy (1948) which compared the social adjustment of educable retardates to normals in adulthood. The consen sus was that the retardates tend to have normal family lives, obtain employment in unskilled and semi-skilled jobs and to become homeowners, but were the first to be unemployed in times of depression. Regular Versus Special Class Placement for Educable Retardates Underlying the establishment of special classes for the educable mentally retarded is an assumption that in this way the retardate will benefit by learning more and by achieving a better level of personal and social adjustment than he would have in regular classes, and that the 25 students in regular classes would profit by the removal of retardates as a result of additional attention of the teacher which foimerly was re quired for the less capable students. In reviewing the literature, we found no studies which attempted to assess the latter benefit, and the findings of investigators of the relative merits of regular versus special class placement for educable retardates are inconclusive. Dunn (1958) summarized the research on mental retardates up to 1958 in his article in the 1960 Encyclopedia of Educational Research. In terms of social adjustment in the schools, studies by Johnson (1950), Johnson and Kirk (1950), and Miller (1956) indicated that retardates placed in regu lar classes tended to be socially segregated. This appeared to be related to personality differences rather than to academic deficiencies. •i Furthermore, there was a tendency on the part of the retardates to be lieve that they were more socially accepted by the others than they actually were. In reviewing the educational provisions for the retarded, Cain and Levine (1963) divided the training conditions into four categories: (a) special versus regular classroom situations, (b) heterogeneous versus homogenous classrooms of retarded children, (c) school versus non-school situations, and (d) community versus institutional school situations. Thurstone (1960) found that mentally retarded children in special classes were not achieving as well as retarded children in regular classes. This finding held for nearly all of the various age groups in the area of academic achievement, although the children in special classes rated above the regular class retardates on personal-social 26 traits. Stanton and Cassidy (1959) compared EMR's in special and regu lar programs from different school districts and also found the regular class children with greater academic achievement but lesser social achievement. Ainsworth (1959) compared changes in achievement and so cial and emotional development of retardates in special classes, regular classes, and classes served by traveling teachers, and found no signifi cant differences. Wilson (1960) found no differences between special class and regular class retardates (who had been matched for mental age and social age) in their social development, speech response, or word recognition; but the regular class group performed significantly better on the Stanford Achievement Test of word meaning. Mullen and Itkin (1961) were able to find no significant difference between regular and special classes over a two year period in academic achievement, in ad justment, in sociometric ratings, or on a behavioral checklist. Three years after the Cain and Levine article, Blackman and Heintz (1966, 5) outlined the recent literature on special class placement in the Review of Educational Research. They found that the studies continued to be inconclusive. Johnson (1962) concluded that mentally handicapped children enrolled in special classes achieved significantly less than comparable children who remained in regular classes, despite small class enrollments, high educational costs, and specially trained teach ers. He attributed the negative findings of the studies primarily to teacher education programs which have stressed both the inability of the retarded and the need for establishing a good mental hygiene program. Kern and Pfaeffle (1962) found that retarded children in special classes 27 showed much better social adjustment, as measured by the California Test of Personality, than did regular class retardates. In contrast, Gold stein, Moss, and Jordan (1965) found that special classes produced higher gains in academic achievement than did regular classes for retar dates, but it is significant to note that their subjects were from the lower part of the EMR range. Sparks and Blackman (1965) pointed to the increasing emphasis on special class placement and special preparation of teachers for the educable mentally retarded despite the lack of em pirical evidence that differences really exist in regard to special class techniques, material, and content. In a similar vein, Simches and Bohn (1963) examined the major curriculum guides for the educable men tally retarded and concluded that the existing programs for the mentally retarded did not differ appreciably from those for the normal classes. One important question which does not seem to be getting much no tice in such studies as those cited above is whether the achievement test instruments used to test regular students can justifiably be used to measure the achievement of the different goals which have been es tablished for the special classes. If present achievement tests are heavily oriented toward the college preparatory type curriculum, perhaps they do not have much value for students who are presumed not to have college potential. Cultural Differences in Intelligence Available studies indicate a positive relationship between intel ligence and occupational status. Stewart (1947, 5-41) did a large scale analysis of the A.G.C.T. scores and occupations of enlisted men in World 28 War II and found that the test scores of groups from different levels of the occupational hierarchy showed large and statistically significant differences (Anastasi, 1958, 515-516). A later study using the pre decessor of the WAIS (Simon and Levitt, 1950, 23-25) achieved similar results with a large sample of civilians in metropolitan New York. Miner (1956) found significant differences between occupational levels in a large sample using a vocabulary test. The correlation between occupational status and intelligence of adults is echoed in the studies of the intelligence of children as re lated to the occupations of their parents. Cronbach (1961, 517) pointed out that this relationship cannot be explained entirely in terms of dif ferences in vocational experience and amount of formal schooling. McNemar (1942, 38) illustrated this relationship in the revision of the Stanford-Binet Scale in which the seven occupational classes (1. Pro fessional; 2. Semiprofessional and managerial; 3. Clerical, skilled trades, and retail business; 4. Rural owners; 5. Semi-skilled, minor clerical, and minor business; 6. Slightly skilled; and 7. Day laborer, urban and rural) clearly reflected a positive relationship between pa ternal occupation and intelligence of the child. Havighurst (1949) obtained significant relationship between social class membership and I.Q. scores on Thurstone's Tests of Primary Mental Abilities. He used children ranging from 10 to 16 years of age and the results were con sistent for all age groups. He found that the high socio-economic status children had their greatest advantage in the area of Verbal Comprehension and the least advantage in memory. 29 Fishman (1964, 127-145) compared the characteristics of middle class and lower class children and concluded that the lower class child tended to be less verbal, more fearful of strangers, less self-confident, less motivated toward scholastic and academic achievement, less compe titive in the intellectual realm, more irritable, less conforming to middle class norms of behavior and conduct, more apt to be bilingual, less exposed to intellectually stimulating materials in the home, less varied in recreational outlets, less knowledgeable about the world out side his immediate neighborhood, and more likely to attend inferior schools. Justman (1967, 354-362) reported a recent study of intelligence test and achievement test scores of disadvantaged children which raised some serious doubt as to the reasonableness of equating low intelligence with low SES. The data in this study were drawn from the record cards of a group of 934 pupils who were attending 16 schools located in dis advantaged areas of New York City. All of the subjects had entered New York City public schools for the first time either in kindergarten or in the first grade, had had uninterrupted attendance in the regular grades of the New York City public schools, and in each case there was an Otis Alpha Intelligence Test score as a result of testing at the third grade level and there was an Otis Beta Intelligence Test score from testing in the latter part of the sixth grade. Of the total group 42 per cent had attended only one school. The others had attended two or more schools during the first six years. When the data were examined, virtually no change was noted in I.Q. for the total group of 934 pupils. 30 However, when the total group was divided into two subgroups, a stable group that had been in a single elementary school for the entire six years, and a mobile group that had attended more than one school, sharp differences in I.Q. were observed. In the stable subgroup there was an increase in I.Q. between the third and sixth grades, and in the mobile subgroup there was a small decrease during the same period. Justman reported that the difference between the two groups was significant. Levinson (1961, 69-82) found that the stability of I.Q. in a par ticular subculture will depend in part on the nature of the tasks which are considered as indicators of intelligence. If the tasks measure the indirect expression of the values of a given subculture and there is no change in these values, the I.Q. will remain stable. If there is an in creasing emphasis in school or elsewhere on those values which are indirectly measured by the tasks in question, the I.Q. will rise. If there is a de-emphasis on these values, the I.Q. will decline. Anastasi (1958, 522) summarized the results of studies of the in telligence of various isolated groups. One of the best known of these was that of Gordon (1923) who investigated the educational achievement of English "canal-boat” children. Their formal education consisted of rare attendance at classes conducted on special boats which were tied to the docks so that the children could attend school while the boats were being loaded and unloaded. Gordon indicated that this amounted to about five percent attendance for these children. In addition the home en vironment had little intellectual stimulation with many of the adults being illiterate, and there was little opportunity for contacts with 31 individuals outside the family. Average I.Q. for these children was found to be 69, but more important, there was a high negative correla tion between I.Q. and age, indicating that as the children got older the impoverished cultural environment resulted in gradually lowering intelligence as compared to the normal population. By contrast, Gordon described another relatively isolated group whom he referred to as the "gipsy children." This group had somewhat more educational opportunity, was less isolated from outside influences and experiences, and the re duction in I. Q. with increasing age was much less pronounced. Anastasi (1958, 524) presented a number of studies of isolated children in the United States (Asher, 1935, 480-486; Hirsch, 1928, 183-244; Chapinis, 1945, 27-55; Wheeler, 1942, 321-334; Edwards, 1938, 317-333; and Sherman, 1932, 279-290). These groups were characterized by a low standard of living and a great deal of inbreeding. In all cases the average I.Q. was considerably below the national norms with Verbal being lower than Performance Scales and I.Q,. declining with age. There was a direct relationship between the amount of mental retardation and the degree of isolation. Anastasi (1950, 13-17) has shown that rural children tend to have mean I.Q.'s significantly lower than city children. Children in the suburbs also have I.Q.'s somewhat lower than those of city children, but the differences do not reach significant levels. Generally the Verbal I.Q. differences are greater than the Per formance I.Q. differences. Rural children also appear to be more handicapped on timed tests than on untimed tests (Pressey, 1920, 91-96). Angelino and Shedd have found that the differences between city and 32 rural children on intelligence tests have been decreasing, probably due to such causes as the recent changes in communication, education, trans portation, and mobility in the rural areas. Performance on intelligence tests has been found to be highly cor related with amount of education as measured by total number of years of school attendance. Yerkes (1921, 15) demonstrated this in a large scale study of service men in World War I, and Fulk (1952, 34-35) obtained similar results in his study of the relationship between Army test scores and last school grade of the service men in World War II. Fulk found that this trend applied to both white and Negro servicemen. When he compared the intelligence of the Negroes and the whites holding last grade in school constant, the whites still excelled the Negroes signifi cantly. This he attributed to the differences in the quality of educa tional opportunity for the two groups. Anastasi (1958, 556) pointed out that even where effects have been made to hold educational opportunity constant by using white and Negro subjects attending the sane schools (Tanser, 1939) there still remained significant differences in the so cioeconomic level of the two groups. Tanser reported that the white children attended school more regularly than the Negro children. So far, efforts to hold educational and socioeconomic opportunities con stant in racial comparisons on intelligence have been unsuccessful. Racial Differences in Intelligence There is a vast literature comparing various group categories in terms of intelligence. The usual procedure for studying racial differ ences in intelligence involves selecting two racial groups, for example 33 white and non-white, obtaining a number of subjects from each group, controlling age, sex, and perhaps another variable or two such as resi dence, and then administering a standardized (on white subjects) intelligence test. Mean I.Q.’s for each group are computed and the data are subjected to statistical analysis to see if the differences are sig nificant. In such studies nearly always the differences are in favor of the whites, and usually the differences are considered significant as in the following typical studies which found significant white super iority over Negroes: Brigham (1923), Pintner (1931), Boynton (1933), Pasamanick (1956), Carson (1960), Deutsch and Brown (1964), Stewart (1964), Levinson (1964), Shuey (1958), Garrett (1947), Roen (1960), Semler (1966), Lacy (1926), Caldwell (1954), Koch (1926), Moore (1942), Bruce (1940), Swegler (1920), Boots (1926), Cavins (1939), Peterson (1929), and Collom (1937). No studies were found which showed Negro superiority over white subjects; however, some studies have been pub lished which found Negro I.Q.'s in the highly gifted range. The question of Negro-white differences in intelligence has been under study for a long time. One important question has been whether the differences are the same at all ages. McGraw (1931, 1-105) pub lished an important study which found white infants superior to Negro infants on the Buhler Babytests which tend to emphasize social behavior and problem solving to a greater extent than do most infant scales. One aspect of McGraw's study which has been criticized was the fact that the white babies were taller and heavier than the Negroes, which suggests that the two samples were not equal in terms of prenatal care and 34 nutrition (Anastasi, 1958, 582). More recently, Pasamanick and Knoblock (1955, 401-402) tested a group of Negro infants on the Yale Development Examination and found them equal to the white norms for the test. A follow-up study two years later still showed no inferiority on the part of the Negro children. A study of Negro children by Williams and Scott (1953, 103-121) found that lower class children scored significantly higher than middle class children in the motor area. Comparisons of Negro and white infants by Gilliland (1951, 271-273) using the Cattell Infant Intelligence Scale supported the findings of Pasamanick and Knoblock. Anastasi (1952, 147-165) compared Negro and white language devel opment and I.Q. using measures of linguistic development and the Goodenough Draw-a-Man Test. The subjects were 100 five-year-old Negro and white children attending the New York Department of Welfare Day Care Centers, and included 25 Negro and 25 white living in uni-racial, un mixed neighborhoods and 25 Negro and 25 white living in interracial, mixed neighborhoods. The sex ratio was relatively uniform and no sig nificant differences were found in the Draw-a-Man, but the girls excelled the boys for the sample as a whole. Mean sentence length yielded a significant race-by-sex interaction. In the white group, the girls surpassed the boys, while in the Negro groups, the boys excelled. This reversal of sex differences among the Negroes was more pronounced in the unmixed neighborhoods. Race differences in mean sentence length favored whites but was slight and not significant. Whites tended to use more mature sentence types than did the Negroes. 35 Green (1967, 5-14) contended that among the factors in, a child's immediate context studied as germane to its intellectual growth are socio-economic class, parental education and marital status, number of siblings, presence of a radio in the home, the amount of prior education, and grade level. Not surprising were the results of five studies (Arlitt, 1922, 378-284; Horton, 1962, 255-265; John, 1962; Banks, 1955, 1200-1201; and Kennedy, 1961) which showed positive relationship between increase in intelligence and increase in socio-economic status of Negroes. Drawing attention to the influence of SES on intelligence was the study by Jenkins (1943, 159-166) of Negro children having Stanford- Binet I.Q.'s of 160 and above, none of whom had parents in lower than the middle class SES. The lowest educational level among the parents of these gifted children was high school graduation. Green (1967) found that for Negro subjects intelligence increased with decreasing number of siblings and decreasing frequency of unmarried parents. Lorge (1945, 483-492) tested intelligence before and after a twenty year in terval and found change in I.Q. to depend greatly on the education received in between. A group of investigators (Green et al.» 1964) reported a study in depth of children in a Southern county whose schools had been closed for several years to prevent integration. The predominantly low educa tional level of the parents offered little alternate instruction at home. Intelligence, achievement, social, and demographic data were col lected for about 1700 children aged 5 to 22. The sample was divided into those having had no schooling whatsoever and those receiving at 36 least some out-of-county education. The mean I.Q. of the education group was approximately 80. Up to the age of eight, this was also the mean I.Q. of the no-education group; after age eight, the overall mean I.Q. of children without any formal schooling averaged 65. I.Q. for both groups appeared to be lower, the older the age group tested. Thus, in an actual sample of children undergoing the nonschooling treatment, measured intelligence level appeared to be dependent upon formal in struction at all ages. There are fewer white and Mexican than white and Negro intelli gence comparisons in the literature, but the results were essentially the same--all in favor of the whites, as for example: Sabagh, Dingham, Tarj an, and Wright (1959), Altus (1953), Silverstein (1962), Garth (1928), Garth (1936), Garretson (1928), Sheldon (1924), Garth and Johnson (1934), Goodenough (1926), Hanson (1931), Paschall and Sullivan (1925), Manuel (1932), Koch (1926), and Davenport (1932). Palomares and Johnson (1966, 27-29) have pointed out that in some school districts in Southern California, Mexican-American pupils form 85 percent of the enrollment in special classes for the educable men tally retarded. These investigators have reservations concerning the appropriateness of the WISC for identification of retardates in the case of Mexican-American pupils, especially when the examiner does not speak Spanish. Their sample consisted of 50 pupils, 30 males and 20 females, ranging in age from seven to fourteen; mean age was 11.2. Their grade placement ranged from 1.0 to 7.0, with a mean grade place ment of 4.2. The subjects were given the WISC, the Goodenough Draw-a- 37 Man Intelligence Test, and the Wide Range Achievement Test. Mr. Palo- mares speaks fluent Spanish and has a background similar to that of the Mexican-American subjects, while Dr. Johnson does not speak Spanish. As additional test scores to be used in the comparison^ the Leiter In ternational Performance Scale was administered to all fifty subjects in their English classes. Analysis of their data revealed that of the fifty Mexican-American pupils who had been referred for EMR testing, 38, 12, 28, and 6 were found to be eligible for EMR classes on the basis of the Verbal, Performance, Full Scale (WISC), and Goodenough I.Q.’s respectively. The investigators computed achievement discrepancy scores based on the distance below grade level the student scored on the Wide Range Achievement Test. The discrepancy achievement scores were found to be unrelated to any of the I.Q.’s. The lack of relation was interpreted by the investigators as indicating that, for this sample at least, I.Q. was not the most relevant determiner of whether a pupil is achieving up to grade level. They did find a high relationship be tween age and discrepancy score with the older pupils having the higher discrepancy scores. A comparison of EMR recommendations by the two examiners was made from a sample of 68 pupils screened for EMR placement--33 by the non-Spanish speaking psychologist and 35 by the Spanish-speaking psychologist, with both psychologists using the WISC. Both examiners gave all test instructions in English. There was a written report on every subject. The non-Spanish speaking psychologist found 24 of his 33 (73 percent) eligible for EMR placement, but in con trast, the Spanish-speaking psychologist recommended only nine of his 38 35 cases (26 per cent) for EMR placement. The investigators attributed this large difference to the language factor. However, this aspect would need further study before such a conclusion would be justified. Certainly there are other possible explanations for these findings. < The investigators suggested that the Leiter International Performance Scale and the Goodenough Draw-a-Man Intelligence Tests should receive more use by those evaluating Mexican-American pupils. Holland (1960) investigated the "language barrier" as an educa tional problem of Spanish-speaking children. He gave WISC's to 36 Spanish-speaking children recommended for testing in the Tucson public schools. The WISC instructions were translated into Spanish. Students were given the test in English and the responses were scored. However, if the student did not understand an item, it was administered a second time in Spanish and scored additionally. (This was not done in the Digit Span and the Performance subtests.) The- system yielded two Verbal scores: an English Verbal I.Q. and a bilingual I.Q. The first repre sented the present level of functioning in English language skills, and could be considered to indicate most clearly the present level of lan guage achievement in the school situation. The second suggested the future potential for Verbal skills when the subject's knowledge of English would be equal to that of Spanish. The I.Q. point difference between these scores was referred to as the language barrier. The in vestigators reached the following conclusions: (1) A child has a language barrier when his knowledge of Spanish is greater than his knowledge of English. This condition originates in lack of acculturation. 39 (2) There was an average language barrier of 4.6 I.Q. points per stu dent. In eight cases the language barrier was very serious, in seven cases the language barrier was regarded as serious, and in eighteen cases it was moderate. Only three students were completely free of the lan guage barrier. (3) Although language barrier declines with increased schooling, it is still present among some fifth graders. (4) Although grade 1C (an additional year of preparation for first grade level ex periences placed between kindergarten and first grade) reduces language barriers considerably, many students beyond this level still have seri ous language problems. (5) Language barrier is a very important factor in the lower academic achievement of many Spanish-speaking chil dren recommended for testing. (6) Aside from the language barrier, the substandard Verbal development of the Spanish-speaking children is probably the result of being bilingual and having to forfeit a more thorough knowledge of one language for a partial familiarty with two. Hernandez (1967, 59-65) maintains that the greatest limitation facing today's teacher of the Mexican-American is lack of knowledge of the student's culture and background. Teachers function daily in the classroom and on the playground with a concept of their students and of their student's parents which can only be described as stereotyping. The stereotype is all-inclusive: Mexican-Americans are seen as unde pendable, irresponsible, indolent, dirty, and unhealthy. They have questionable moral standards. They refuse to learn English and stub bornly cling to their native language and Mexican culture. They refuse to help themselves and ignore educational opportunities. In 40 the classroom they are perceived as taciturn or laconic. They never study or do homework. They always come to class unprepared, without books, without pencils. It is impossible to teach a continuous lesson because they are always tardy or absent. The majority of them are not interested in grades. Moreover neither the parents nor the children j jtake an interest in the school or the community. Hernandez suggests jthat the Mexican-American student is caught in a trap between conflict ing demands and cultural expectations of the home and the school. | Increasing the frustration experienced by many Mexican-American Jstudents is the tendency to fall farther and farther behind as they go up the grades. Garth and Johnson (1934, 222-229) made a survey in the I schools of El Paso, Texas, and the rural schools of New Mexico which in-! eluded 683 Mexican children. They found that in both achievement and ' intelligence the Mexican children started out much like the American whites but became less and less like them age for age as the age level increased. In contrast to the implications of Garth and Johnson's study; is that of Davenport (1932, 304-306) whose data suggested that Mexican children tended to gain in I.Q. as they went up the grades. Description of the WISC and the WAIS The WISC was first published in 1949 and grew logically out of the; Wechsler-Bellevue Intelligence Scales used with adolescents and adults. Most of the items in the WISC are from Form 2 of the earlier scales, the; main additions being new items at the easier end of each subtest to permit examination of children as young as five years of age. The WAIS was first published in 1955, six years after the appear ance of the WISC. It is an extension and modification of the Wechsler- Bellevue Form 1 and is intended for use with subjects age 16 years and older. The norms for the two tests do not overlap, but together they account for subjects of all ages from five years up. The Wechsler i f jScales differ from previous intelligence scales in three principal ways:j ! j i(l) The complete renunciation of the concept of mental age in favor of j jthe deviation I.Q. concept. (2) An extended definition of intelligence jin which it is an inseparable part of a larger whole, namely, person- | lality itself. (3) The subdivision of the tests into two subscales, j i Identified as Verbal and Performance. The Verbal subtests are: In- j j ; formation, Comprehension, Arithmetic, Similarities, Vocabulary, and Digit Span. The Performance subtests are Picture Completion, Picture Arrangement, Block Design, plus a fifth subtest which is called "Coding'’ in the WISC and "Digit Symbol" in the WAIS. The WISC has a sixth Per formance subtest called Maze which has no counterpart in the WAIS. The | Maze subtest is considered to be an alternate in case one of the regular jsubtests is found to be inappropriate for a particular subject or in the event that an administrative slip or irregularity renders a subtest ;invalid. i Standardization of the WISC and the WAIS j ; The WISC was standardized on an all white sample of 100 boys and \ 100 girls at each age from five through fifteen years. Fifty-five of i the twenty-two hundred subjects were feebleminded. The sample was 42 based on four geographic areas, distributed in urban-rural proportions ' * * on the basis of the U. S. 1940 population, and the occupations of the children's fathers were similar to all employed white males in the United States at that time. The manual (Wechsler, 1949) contains data on intercorrelations, reliability coefficients, and standard errors for i ! J ! jail subtests and totals. According to Delp (1950) the primary advan- t ! Stages of the WISC include its modem construction, ease in administrar j |tion, interesting materials for children, convenient manual with clear j directions and tables, well-developed norms, Verbal, Performance, and Full Scale I.Q.'s directly comparable for the various ages, and future possibilities for clinical diagnosis from both quantitative and quali tative points of view. Delp also points out some disadvantages: too i • ! : j few items easy enough for lower age brackets, lack of mental ages which i many clinicians may want, not very useful for subjects at either ex tremes of the intelligence distribution, considerable subjectivity 'in scoring some of the Verbal items, dubious evidence of validity, validity apparently based on assumption of comparability with adult scales, no t |substantiation in manual or elsewhere of implied values of interpreta tion of subtests. As in the case of the WISC, considerable care was taken in the standardization of the WAIS. Norms were developed for each of seven t age groups ranging from 16 to 64 years using 1700 subjects in the norm j f * i I group. I.Q.'s were developed separately for each group. Equal numbers ;of men and women were included in each age group. The United States was divided into the four major geographical regions as in the census 43 reports and cases were assigned to regions in proportion to their popu lation. Urban-rural subjects and white and non-white subjects were represented in the same ratios as in the 1950 census. Thirteen occu pational categories were also established on the basis of census groupings. i j The standardization of the WISC and the WAIS appear to be com- i • parable except for the fact that the WAIS norm group included non-whites. I and the WISC norm group did not. This fact has been suggested as a jpossible reason for differences in I.Q. scores on the two tests j(Webb, 1964). I I I I i I ! | WISC Versus WAIS Reliability ! Wechsler provides tables of reliability coefficients for the Ver- j bal, Performance, and Full Scales, as well as for each of the subtests, ; in the test manuals. In each case, with the exception of Digit Span, split-half reliability coefficients were obtained for every subtest by j 'computing the correlation between scores on odd and even items and j j I ! correcting the coefficient for full length of the test by the Spearman- | 1 r j IBrown formula. An inspection of the tables in the two manuals reveals j that the WAIS subtests have considerably higher reliabilities than those I • i j ! jof the WISC. This might be expected in view of Bloom's (1964) work | ! ' i jwhich established the fact that throughout childhood stability in in- | j * | jtelligence is related to age. It is noteworthy that the Vocabulary sub-] I ' | test has the highest reliability of all the subtests on both the WISC j and the WAIS at all ages. The few test-retest reliability studies done' :so far have yielded very satisfactory results, as for example, Bayley (1957), Armitage and Pearl (1958). Guertin, Rabin, Frank, and Ladd j i j (1962), surveying the preceding five years of research on the WAIS ' i found that test-retest differences were not only greater, but more var iable for the Performance I.Q. than for Verbal or Full Scale I.Q.'s. At the subtest level, their survey revealed that the test-retest reli abilities are "generally higher" than the split-half reliabilities reported in the WAIS manual. Practice effects were found to be much more variable on the Performance subtests than on the Verbal subtests. WISC-WAIS Validity Studies j One of the more perplexing problems in the field of mental test- | iing is that of establishing the validity of the testing instrument. In| ! ' * I I | reviewing the research on the WISC for the period between 1949 and 1959] Littell (1960) pointed to the lack of any explicitly stated, organized : I network of intuitive reasons for expecting it (the WISC) to show pre dictive validity other than the very broad assumption of a . general | factor which enters into the purposeful solution of all problems. I ; jWechsler asserts that intelligence cannot be viewed separately from I I personality and that non-intellective factors enter significantly into ' ;a child’s actual behavior in problem situations, yet the Wechsler testsj ; i ! do not sample these factors in any systematic manner. Littell was unable : to find any predictive studies at all in the literature concerning the j | WISC; rather the emphasis was on the comparability of scores on the ; I WISC and other instruments for various groups. Frandsen and Higginson , (1951, 283) reported a study on 54 fourth grade children and concluded that the Binet and the WISC yield comparable scores at least within 45 one or two standard deviations above and below the mean. Pastovic and Guthrie (1951, 385) summarized five studies and recommended that the WISC not be considered comparable to the Stanford-Binet below the age of ten since the WISC scores were considerably lower than the Binet for those ages. Krugman, Justman, Wrightstone, and Krugman (1951, 482) found that Stanford-Binet scores were higher than WISC scores to a sig nificant degree only at the younger ages. Weider, Noller, and Schraumm (1951, 332) found that while the Stanford-Binet and WISC are signifi cantly correlated, the Binet I.Q.'s tend to be higher than the WISC I.Q.'s for the same children. Other researchers (Kureth, Muhry. and Weisgerber, 1952; Levinson, 1959; and Triggs and Cartee, 1953) also re ported that children within noimal I.Q. ranges tended to have higher Stanford-Binet's than WISC's. Many studies have reported that the WISC Verbal Scale correlates higher than the Performance Scale with the Stanford-Binet, but none was found where the reverse was true. Guertin, Rabin, Frank, and Ladd (1962) in their five year review of the literature concerning WAIS studies expressed disappointment at the absence of validity studies which dealt with the instrument's pre dictive accuracy. They found that studies reported for the period between 1955 and 1960 tended to ignore problems of restricted range. Tarj an et at. found that the release rate of male and female institutionalized mental defectives was predicted by the Full Scale WAIS I.Q. (1960) However, another team of researchers (Miller et al.3 1961) found that release rate of male retardates who were below age 30 was predicted by the Verbal I.Q. but not by the Perfoimance I.Q. 46 Neither scale predicted the release of females. A third study (Appell et al.3 1962) compared WAIS I.Q.’s with suc cess in adjusting to outside job assignments by mental defectives and the authors reported that those succeeding tended to have significantly higher Performance than Verbal I.Q.’s. Baumeister (1964) found no relationship between the WISC Full Scale I.Q. and success in a paired-associates learning task. In a review of the comparability of intelligence quotients of men tal defectives on the WAIS and the 1960 revision of the Stanford-Binet, Fisher, Kilman, and Shotwell (1961) found that the correlation coeffi cients between I.Q.'s of the two tests tended to run between .60 and .90 in noimal, psychoneurotic, and mentally retarded populations. There was some tendency to get higher WISC than Stanford-Binet I.Q.'s but the two instruments were found to be fairly comparable. Wechsler (1958) found the mean Stanford-Binet I.Q. to be five points higher than the WAIS with a group of 52 prisoners aged 16 to 26 years, and the correla tion between the two tests was .85. Fisher, Kilman, and Shotwell (1961) studied the effect of age and level of retardation of the comparability of I.Q.'s from the WAIS and the 1960 Stanford-Binet. They found that age, but not level of retardation, was significant in determining the magnitude of the difference between WAIS and Stanford-Binet I.Q.'s- WAIS averaged 15 to 23 points higher than Stanford-Binet for subjects 18-54 and 55-73 years respectively. It should be kept in mind that these were institutionalized retardates rather than educable retardates en rolled in public school classes. In a later comment about this study 47 Fisher (1962) pointed out that for subjects tested with the Stanford- Binet in their formative years (6c'14) and then later retested at the peak of their intellectual efficiency (19-34), with the Stanford-Binet and the WAIS, there is constancy for the Binet I.Q.'s whereas WAIS I.Q.'s are significantly higher than the Stanford-Binet's. Subjects who are tested with the Stanford-Binet in their peak years give fairly constant Stanford-Binet I.Q.'s when retested during later years (35-70), whereas their later WAIS I.Q.'s are all significantly higher. Fisher found a tendency for WAIS I.Q.'s to become increasingly larger than Stanford-Binet's as age increases, and he concluded from this that for mental retardates the Stanford-Binet gives a reliable measure of intel ligence over a long period of time and that the WAIS I.Q. 's are arti- factually affected by age and therefore have questionable validity. He suggested that the norms for the WAIS may be inadequate in that there was insufficient sampling of subjects of lower intellect and, as a re sult, the rate of decline of intellectual abilities with age for the standardization sample is inapplicable to mental retardates. In the same year Fisher (1962-b) reported a longitudinal study of institutional retardates comparing WISC I.Q.'s with later WAIS I.Q.'s on the same individuals. This was the first WISC-WAIS comparison using the same subjects for both tests to be found in the literature. Fisher noted that the discrepancies between WISC and WAIS were of about the same size and were of the same direction as he had found between Stanford-Binet and WAIS I.Q.'s. From this he concluded that the WISC should be considered the more valid of the two Wechsler tests for use 48 with retardates. The second WISC-WAIS comparison was by Webb (1963) of the Pasadena City Schools. As part of a routine re-examination of students enrolled in special classes for educable mentally retarded, the WAIS was adminis tered to 20 Negro subjects each of whom had a WISC examination approxi mately two years earlier. The ages of the subjects at the time of the earlier (WISC) testing ranged from 13.1 to 15.8 with a mean of 15.2 years. The ages of the subjects on the WAIS testing ranged from 16.0 to 18.5 with a mean age of 17.1. Webb found no sex differences in his study, but he had a very small sample, twelve girls and eight boys. The range of Full Scale I.Q.'s for the WISC was 53 to 80, and for the WAIS, 68 to 94. He found significant differences between WISC and WAIS for all three scales: Verbal (p=.001), Performance (p=.02), and Full Scale (p=.001), with the WAIS being the higher in each case. He also noted that in spite of the differences in I.Q.'s, there were between-test correlations for all three scales: Verbal Scales, r=.80; Performance Scales, r=.91; Full Scale, r=.84. Even the subtest correlations be tween the two tests were high, and this was in spite of the fact that the WAIS Digit Symbol and Picture Arrangement subtests were signifi cantly higher (p=.10) than their WISC equivalent. Webb viewed his most impressive finding as being the significantly higher WAIS I.Q.'s, and he suggested that this might be due to differences in standardization populations for the two tests. The WAIS included non-whites in the standardization group whereas the WISC did not, and Webb reasoned that the WAIS might therefore be the more valid test for use with subjects 49 as his. Webb followed up his 1963 study with another (Webb, 1964) in which he investigated race and sex influences on WISC-WAIS differences. He selected two groups this time, one Caucasian and the other California- born Negroes, who had been tested with the WISC in the eighth grade and with the WAIS in the eleventh grade. All of the subjects had been en rolled in classes for educable mentally retarded for at least three years. He matched the Caucasian and Negro subjects on the basis of sex and length of time in EMR classes. There were nine pairs of girls and seven pairs of boys for an N of 32. In this study, as in the previous study, he found no significant sex differences, so he disregarded sex as a variable for the rest of the analysis. He used a 2 x 3 analysis of variance design to test race and test differences and the interaction between test and race groups. For Full Scale I.Q.'s there were no ethnic differences found, and the interaction effect was not signifi cant. However, differences between WISC and WAIS I.Q. scores were highly significant (p=.01) for all three scales in all cases in favor of higher WAIS. Race and interaction effects in both cases were not sig nificant. There was a non-significant tendency for the Caucasians to score higher on all three scales of both tests. Webb concluded that his findings were very much like those of his previous (1963) study and like those of Fisher (1962-b) and that there was little doubt that the WAIS yielded higher estimates with mental retardates. This phenomenon was found to apply equally to Caucasians as to Negroes. Since in all cases the WISC was administered first and the WAIS second, Webb applied 50 a regression equation (y'=rx) to his data to determine whether a re gression effect could account for the higher MIS scores. For the Negro EMR's with an average WISC I.Q. of 66.68, the expected average WAIS I.Q. on the ba sis of regression would be 72.70 where the actual average WAIS I.Q. was 76.12. For the Caucasian EMR's with an average WISC I.Q. of 70.00, the ex pected average WAIS would be 75,10. The actual average was 80.93 (Webb, 1963). » • Webb concluded that regression accounted for at least part of the higher MIS scores, but still leaving.significant differences between the two tests. There are several points of interest in comparing the work of Fisher and Webb. Both worked with retardates and both found consistent significant I.Q. differences in favor of the WAIS. However, Fisher's subjects were institutionalized retardates while Webb's subjects were educable mental retardates in public schools, considerably higher in average intelligence than Fisher's subjects. Fisher based his prefer ence for the WISC on the fact that the WISC-MIS comparisons of his second study yielded results similar to the Stanford-Binet-MIS compari sons of his first study. He had found the Binet to demonstrate a high level of reliability over time for his institutionalized retardates and from this he believed that the WISC might therefore have validity as well. Since the age norms of the MIS do not apply to subjects younger than 16, no comparisons of long-term reliability for the WAIS can be made for subjects in what Fisher described as the "formative years of 6 to 14." The disagreement between Fisher and Webb as to which of the two Wechsler tests is the more "valid" for mental retardates remains 51 unresolved perhaps because of the absence of an adequate operational definition of the validity that is involved. Both Fisher and Webb ap pear to be limiting their immediate consideration to concurrent validity with Fisher using the long-established Stanford-Binet Scale as the cri terion with which to compare the two Wechsler Scales and with Webb using other clinical data: Using the criteria of social history, developmental history, family history, and physical examination, the students who no longer qualify for EMR placement on the basis of WAIS scores also do not appear men tally retarded on these variables. Teacher estimates of the students seem to confirm these observations. (Webb, 1964). One of the objectives of the present study will be to assess the relative effectiveness of the WISC, WAIS, and individual reading tests in predicting probability of success in regular classes for below- average intelligence students. Subtests of the WISC and the WAIS Information Infoimation tests measure what the subject already ]<nows as well as his cultural background. Memory and alertness to the environment are important aspects which will determine how well the subject does on this test. Wechsler (1958, 64-65) pointed out that psychiatrists have long used questions calculated to tap a subject's range of information and to estimate their intellectual level, but psychologists have tended to exclude information type items when devising intelligence scales. It was in the development of group tests that information questions became accepted and the first major support for range-of-information as a good 52 measure of intelligence was provided by evaluation of the data from the Army Alpha Examination of World War I. Correlational analysis showed that the information test was much better than some of the previously highly esteemed types, as for example the test of arithmetic reasoning. Comprehension According to Wechsler (1958, 68) the Comprehension subtest is a measurement of practical information and. general ability to evaluate past experience. The questions, for the most part, are concerned with problems that the individual would likely have met himself, observed others encountering them, or have read about. Wechsler deliberately selected items which had stereotyped answers, involved no unusual words so that individuals with limited education would be able to understand. However, poor verbalizers still tend to do poorly on the Comprehension subtest. It correlates best with Information and Vocabulary and least well with Digit Span and Object Assembly. Kitzinger and Blumberg (1951) pointed out that the moral values of the subject tend to influence an swers to the Comprehension test. The earlier items required little or no abstract reasoning. Plumb and Charles (1955) studied scoring dis agreements to the Comprehension responses and found that experts as well as graduate students disagreed significantly. The manual provides sample answers but Wechsler admits that it would not be feasible to in clude every possible answer and he suggests that the examiner use his own judgment when there is an unusual answer that does not fit the models. 53 Arithmetic According to Wechsler (1958, 69) the ability to solve arithmetic problems is one of the oldest recognized signs of mental alertness. Most intelligence tests include some items of this type. Although it correlates highly with general intelligence measures, it is vulnerable to influence of educational and occupational background to a greater extent than are most of the other measures. Clerks and businessmen usually do well but housewives and laborers usually do poorly on it. Adults are less likely to be embarrassed by the arithmetic problems than they are by such chiId-identified tasks as the Block Design. The cor relations between the Arithmetic subtest and the total I.Q. are neither the highest nor the lowest obtained. Kitzinger and Blumberg (1951) point to the fact that the Arith metic subtest emphasizes concentrated attention and mental effort. The arithmetic items are not identical in the mental processes involved. The earlier items are simple and depend upon school learning, but the later items are more difficult and necessitate an analysis of the prob lem using logical reasoning and abstraction. Davidson (1950, 489-492) contrasted the tendency of the Arithmetic subtest to measure active con centration as opposed to the Digit Span which involves the capacity for passive attention and concentration on outer environmental stimuli. The investigator has noticed that low SES subjects tend to have difficulty with some of the Arithmetic items, apparently because they are not asked in the language which the subject is used to hearing. 54 Similarities Wechsler (1958, 72-74) believes that the criticism that the Simi larities item is greatly influenced by language and word knowledge is only partly justified, because it lays heavy stress on the individual's ability to perceive the common elements of the teims that he has been asked to compare and, at higher levels, the ability to categorize them in a single classification. Some of the advantages of the Similarities subtest items are that they are easy to administer, require no extra equipment, and are interesting to the average child and adult. It has particular appeal to the clinician who is concerned with getting more than just an I.Q. from the test administration, because it provides op portunity for qualitative differences in responses and response modes. Correlations for Similarities with Full Scale I.Q. are among the highest of all the subtests. The WISC Similarities subtest consists of 16 items, the first four of which are analogies and are administered only to subjects under the age of eight years and older suspected retardates. Items five through 16 consist of asking the subject in what way two concepts (as for example plum and peach or cat and mouse) are alike. Wiener (1957) found that highly distrustful subjects tend to do poorly on the Similarities subtest as a result of denying that the paired concepts were alike. The WAIS Similarities subtest consists of 13 pairs of items very much like the WISC except it does not have the analogies and four con secutive errors instead of three are required to stop the test. Kit- 55 zinger and Blumberg (1951) have pointed out that while the early items on the WAIS Similarities may be answered easily because they provoke previously made associations and are therefore similar to the Informa tion items, the later items do require judgment as well as the use of the associative processes in the evaluation of essential likenesses. Vocabulary E. L. Thorndike maintained that intelligence and vocabulary were almost synonymous and the emphasis of his efforts to find ways of measuring intelligence was directed at finding out how many words the subject knew and could use meaningfully. Wechsler (1958, 84-85) stated that the size of a person's vocabulary was not only an index of his ed ucation but also an excellent measure of his general intelligence. All four of the Wechsler scales and the several versions and editions of the Binet Scales have featured the Vocabulary test. The major objection raised against the Vocabulary test has been that it is greatly influenced by educational and cultural opportunities. In deference to this objection, the Vocabulary test was used only as an alternate in the early development of the first Wechsler-Bellevue Scale, but its general merit soon became so apparent that in the 1941 edition of the Measurement of Adult Intelligence its use as a regular test was strongly recommended. In the case of the WAIS, the Vocabulary subtest has formed an integral part of the scale from the beginning. In estimating the size of a person's vocabulary the important , factor is the number of words known to the subject rather than the ele gance of his style in defining them. Almost any recognized meaning 56 is accepted (Wechsler, 1958, 84). The Vocabulary test correlates with the Full Scale I.Q. better than any other subtest for all ages on both scales. It was selected by Rapaport (1945, 573) as the individual's hypothetical original level of intellectual functioning, from which to measure impairment in the case of brain damaged subjects. The WISC Vocabulary subtest consists of 40 words of increasing difficulty for which the subject is asked to give the meaning. A printed list is provided for the subject to look at as he takes this subtest. Digit Span The Digit Span Subtest is one of the most popular and most widely used of all the scales of intelligence. It was part of the original Simon-Binet Scale and has been retained in all its revisions. It is easy to administer, easy to score, requires no equipment, and appears to be specific in terms of the ability it is measuring. In spite of these advantages it generally correlates poorest with other tests of intelligence. Wechsler (1958, 70-72) for this reason has contemplated removing it from the WISC and the WAIS, but because of its usefulness as a diagnostic tool, he has retained it as an integral part of the WAIS and as an alternate in the WISC. One unresolved issue which has received a great deal of attention from researchers has been the rela tionship between scores on the Digit Span and .anxiety. Wechsler (1958, 71) maintains that low scores on the Digit Span, when not associated with organicity, can be due to anxiety or inattention. In either case, 57 difficulty in the reproduction of digits correlates with lack of ability to perform tasks requiring concentrated effort. Jurjevich (1963) has found that of all of the WAIS subtests, only Digit Span has a signifi cant relationship with the Taylor Manifest Anxiety Scale or the Welch MMPI A Factor in teenage delinquent girls. However, Guertin et at. (1966, 385-409) reported that the assumption that digit span is influ enced by anxiety or distracting stimuli had been rejected by the studies he had reviewed for the previous five year period. The WISC and WAIS Digit Span subtests are virtually identical ex cept for the numbers used in the seried. In each case the subject is asked to listen carefully to a series of numbers and then to repeat them just as they were given. If an error is made the examiner says another series having the same number of digits. Missing on two series of the same length finishes Digits Forward, whereupon the subject is asked to reverse the order of the numbers in series. Picture Completion In the Picture Completion subtest the subject is shown an incom pletely drawn picture and asked to name the missing part. It is similar to the Mutilated Pictures test of the Binet Scale. Wechsler (1958, 77-78) commented that it was hard to find suitable items for this test. If one chooses familiar subjects, the test is too easy, and if one turns to unfamiliar ones, the test ceases to be a . good test of intelligence because it requires specialized knowledge. The pictures which are included in the WISC and the WAIS were se lected from a much larger group after trying out each one of them over 58 a period of a number of months with a group of subjects with laiown in telligence levels. Each picture was accepted or rejected on the basis of its discriminating value. The major assets of the Picture Completion type test are that it takes little time to administer, practice effects have proven to be minimal after an interval of several months, and it appears to be a good way of testing subjects at the lower intelligence levels. The test requires the subject to perceive the whole in rela tion to its parts, and to differentitate essential from unessential detail. The subject must know what the picture represents and must be able to recognize that the missing part is essential to the form or function of the pictured object. The subject's familiarity with these frequently seen objects may give some indication as to his degree of reality contact (Kitzinger and Blumberg, 1951). Picture Completion correlates higher with Performance than with Verbal tests and usually shows highest loading under the visual motor factor. Picture Completion holds up with age better than any of the other Performance tests, and for this reason it is included as one of the "hold" tests in calculating deterioration ratios (Wechsler, 1958, 78). Wiener (1957) found that highly distrustful subjects tended to get lower scores on Picture Completion as a result of frequent "nothing missing" answers. Picture Arrangement The Picture Arrangement subtest consists of a series of pictures, much like those found in the comics, which, when placed in the correct 59 sequence, tell a story. The pictures are presented to the subject in the incorrect order and he is asked to rearrange them in the correct order. Wechsler (1958, 74-77) outlined the history of this test and pointed out that it was first used by DeCroly, a Frenchman, in 1914. Several Picture Arrangement series were attempted by Army psychologists in 1917 as part of a group battery, but it was not very successful. It was tried by several other experimenters but the problems in scoring and developing good sequences apparently were considered too great. Wechsler and his co-workers tried out a number of cartoon series as items, and after considerable revision and eliminating, Picture Ar rangement was included as part of the regular scales. The sets of pictures were selected on the basis of interest of content, probable appeal to subjects, and ease of scoring and discriminating value, and in spite of considerable labor before the final choices were made, Wechsler felt that much was left to be desired. In spite of its serious weaknesses, Wechsler believes that the Picture Arrangement Subtest has some worthwhile merits, particularly that of measuring a subject's ability to comprehend and size up a total situation. The subject must understand the whole situation before he is able to perform the task. Davidson (1950, 489-492) described the Picture Arrangement Sub test as one which involves the ability for planning and anticipatory action as well as the grasp of social situations and a time element. Kitzinger and Blumberg (1951) describe it as measuring anticipation, planning, and ability to comprehend and size up a total situation in relation to its parts. It is a test of ability to pick out essential 60 cues. It also reflects a subject's capacity for interpersonal rela tionships . The investigator in the present study has been troubled by the frequency with which intelligent subjects "miss" items in this subtest by composing arrangements which do not conform to the acceptable ones, but nevertheless are accompanied by stories which seem to be meritor ious and logically consistent. The WISC Picture Arrangement Subtest consists of three cut-up ink drawings of a dog, a mother and toy scene, and a toy train followed by seven cartoon sequences presented in incorrect order. The test is discontinued after only two failures. The WAIS Picture Completion consists of eight cartoon sequences, the cards for which are presented in the incorrect order as in the case of the WISC. In contrast to the WISC, all of the WAIS Picture Comple tion items are administered regardless of the number of zero scores. Block Design The original Block Design Test was devised by Kohs who intended it as a general measure of non-verbal intelligence. This test has been incorporated into a number of intelligence scales. It correlates well with the total I.Q. and with most of the subtests in both the WISC and the WAIS. It also correlates better with Comprehension, In formation, and Vocabulary than do some of the Verbal tests. Surpris ingly, individuals who do best on Block Design are not necessarily those who see the pattern as a whole,, but are able to break it up into small portions (Wechsler, 1958, 79-81). The investigator of the 61 present study has noted on many occasions that the point at which the subject appears to ’ ’ catch on” to the Block Design is when he approaches it a part at a time (as for example setting the comers separately) rather than trying to work out the whole pattern as a unit. Wechsler believes that the Block Design is not only an excellent test of general intelligence, but that it also lends itself well to qualitative analysis. One can determine a great deal by watching how the subject approaches the task. Wechsler has found the Block Design as the most useful of all the subtests in the study of the effects of organic brain damage. The time allotments are slightly different in the WISC and the WAIS Block Design tests and there may be a tendency for the WISC to be more sensitive to the differences between rapid and casual performance. Obj ect Assembly The Object Assembly Subtest was devised by Pintner and first used on the Pintner-Paterson Scale. Wechsler's form boards used in the WISC and the WAIS are very similar to those used by Pintner. Wechsler re ports that he had a great deal of hesitation about including this test in his adult scale because those used in the past had tended to be much too easy for adults and the high levels had little discriminating ability. Most of them had been found to have low predictive value. In spite of these drawbacks it was included. Object Assembly correlates poorly with most of the individual subtests but it does contribute to the total score. It has received its highest praise from clinical ex aminers because of the information it provided about the thinking and working habits of the subjects (Wechsler, 1958, 82-84). Kitzinger and 62 Blumberg (1951) maintain that the assembling of objects is mainly a test of synthesis--putting things together in a familiar configuration. Suc cess on this test depends upon visual and motor functions. Insights may be obtained as to a subject's persistence, his carefulness or careless ness in working at a task, and the amount of trial and error used in problem solving. This test differs from Arithmetic in that here the ex aminer can observe manifestations of attempted solutions. One of the issues in the research related to the Object Assembly test is whether the test is a measure of ability for cognitive organi zation. Lanfeld and Saunders (1961) correlated performance on the Object Assembly test with other Wechsler tasks and a task similar to the dissected sentence problem of.the Binet. Performance on Object Assembly was more closely associated with that involved with the dissected sen tence problem. On the whole, Lanfeld and Saunders felt that the hypo thesis that Object Assembly was a measure of ability for cognitive organization was only partially supported. The WISC Object Assembly Subtest consists of four form boards which are arranged before the subject in a standardized manner while he is sitting with eyes closed. He is asked to open his eyes and put the form together. Digit Symbol (Coding) The Digit Symbol Subtest is one of the oldest and most established of the various kinds of intelligence tests, and it has been used in a number of individual and group scales. In this test the subject is re quired to associate one set of symbols with another set of symbols, and 63 the speed and accuracy with which he does this is considered to be a measure of his intellectual ability. Wechsler (1958, 81-82) observed that there is some question as to whether speed as well as power should be given weight in the evaluation of intelligence. Since Wechsler be lieves that it should, he has included the Object Assembly in both of his intelligence scales. Research has indicated that unstable indivi duals tend to do poorly on Digit Symbol (Tendler, 1923). Kitzinger and Blumberg (1951) have observed that Digit Symbol is primarily a test of ability to learn an unfamiliar task; it also involves visual motor co ordination and some degree of flexibility. Inability to reverse the letter N may be a sign of rigidity in visual or motor organization. Subtest Contribution to I. Q. One of the objectives of the present study is to determine the extent to which the several subtests of the WISC and the WAIS contri bute to the difference between obtained I.Q. scores on the two tests. No previous research could be found which dealt with this question in relation to educable retardates in the school population, but there were three reports which are pertinent enough to mention. MacPhee, Wright, and Cummings (1947) examined the protocols of 432 rural Negro naval recruits who had been given the Verbal Scale of the Wechsler- Bellevue and whose intellectual fitness for military service had been questioned. The I.Q.'s of this group ranged from 48 to 92 and the average education was 3.09 grades. The authors divided their sample into seven subgroups on I.Q. levels. Analysis showed that for each of the seven levels the typical pattern was one of non-uniformity in 64 subtest contribution. At each level of ability the subtests contri buted unequally toward the total weighted score, and the rank order of the scores on the subtests varied from one level to another. When the rank order of the correlation coefficients of the Negro group was com pared with the rank order of the correlations of Wechsler's standardi zation group, a considerable degree of similarity was indicated. Hirst, (1960) explored the usefulness of a two-way method of an alyzing the subtest scaled scores of the WISC in the diagnosis of re medial reading problems. Using an interesting two-way diagnosis which combined comparisons of subject's subtest performance relative to the population mean and also relative to the subject's own mean, Hirst found that the characteristic pattern for the total reading disability group in general was high Picture Completion and Picture Arrangement and "possibly high" Block Design and Object Assembly. In addition, there was a pattern of low Digit Span, Arithmetic, and Coding, with "possibly low" Vocabulary. Hirst then broke the analysis down into "moderately retarded" readers' and "severely retarded" readers' subtest patterns. Moderately retarded pattern was high Picture Completion and Picture Arrangement with "possibly high" or low Block Design accompanied by low Coding, Arithmetic, and Digit Span. Severely retarded pattern was high Picture Completion, Object Assembly, and Picture Arrangement accompanied by low Arithmetic, Coding, and Digit Span and "possibly low" Vocabulary and Similarities. The investigator professed ignorance as to the meaning or significance of the findings. It would seem that the finding of "either high or low" subtest contributions would be hard 65 to integrate into a very useful theoretical framework. Recently Alper (1967) published the results of an investigation of the subtest contribution to WISC I.Q. scores with institutionalized retardates. It will be recalled that Wechsler set all subtest means at ten and all standard deviations at three. There is an inherent assump tion then that each subtest contributes equally to the total I.Q. score. While this may be true for the population as a whole, it is not neces sarily true for retardates. Alper inspected the WISC protocols of 713 retardates from 15 different contributing institutions in different parts of the United States. His subjects ranged from five years one month to fifteen years eleven months and included 469 males and 244 fe males. He subdivided his sample into four I.Q. levels: 46-55; 56-65; 66-75; and 76 and above. He presented four hypotheses to be tested by his data. The first hypothesis was that the scaled score contributions of the subtests would be unequal, and this was supported. He found that Comprehension and Similarities were consistently higher and Arith metic and Vocabulary were consistently lower among the Verbal subtests. In the Performance subtests, Picture Completion and Object Assembly were consistently higher and Picture Arrangement and Coding were con sistently lower. The second hypothesis was that sex differences would be found, and an analysis of variance confirmed that males scored sig nificantly higher than females on eight of the ten subtests. Only on Coding did females surpass males significantly and on one subtest there were no significant differences. The third hypothesis was that varia bility of subtest scores would be related to I.Q., but it was not 66 supported by the data. The fourth hypothesis was that there would be differences between Verbal and Performance I.Q.'s and an inspection of the scores revealed that the Performance was higher, significant at the .001 level. Diagnostic and Clinical Features Although it was Wechsler's primary purpose to give a global mea sure of intellectual capacity when he developed the four intelligence scales, he hoped that the instruments would also provide diagnostic ma terial useful for the hospital clinician. One of the features of the Wechsler tests is the provision for not only an over-all index of in telligence but also a "Verbal" I.Q. and a "Performance" I. Q. Some quotations from The Measurement and Appraisal of Adult Intelligence (Wechsler, 1958) may best show his point of view in this regard: As regards the diagnostic implications of differ ences between verbal and performance ability as a whole, the findings are that in most mental dis orders impairment of function is generally greater in the Performance than in the Verbal sphere. With occasional exceptions, this holds for psychoses of nearly every type, for organic brain disease and to a lesser degree for psychoneuroses . . . The stan dard deviation of the mean difference between Ver bal and Performance for the normal population is 10.02 . . . In most instances a difference of 15 or more points (between Verbal and Performance I.Q.) may be interpreted as diagnostically signi ficant (155-160). A great many researchers have followed Wechsler's lead in the use of his subtests in pattern analysis which is a statistical attempt to identify diagnostically different groups and individuals on the basis of their differential performance on a number of tests as a battery. 67 Among the procedures denoted as patterning tech niques, perhaps the oldest and the one most com monly used is that of profile analysis. This consists essentially of the graphic representa tion of comparable test scores sequentially ar ranged in which the high and low points indicate high and low test performance.on the tests in question ... A second procedure ... is one in which the scores on the individual tests of a battery are not treated separately but summed up into a hoped-for discriminating total. This method has often been used in constructing apti tude and personality scales ... A third pro cedure ... is factor analysis (163-166). Wechsler stressed the importance of experience of the clinician in using these techniques: ... a very high Similarities along with a very low Picture Completion is definitely indicative of schizophrenia . . . Low Digit Symbol plus low Arithmetic plus poor Digit Span may, other things equal, merely indicate neurotic anxiety; low Digit Span plus low Picture Arrangement and very low Block Design, organicity; low Digit Span plus low Similarities and low Picture Completion pos sible schizophrenia (169-174). Promising as the idea of subtest analysis appeared to be at first, the results to date have been disappointing for the clinician. Littell (1960) pointed to the difficulty in interpreting individual subtest scores when the exact meaning of the subtests is not clear. Balinsky (1949) found that the subtest scales do not measure the same thing in adults at all ages, and Wechsler (1949) noted that the subtests seemed to measure different factors in children than in adults. Littell (1960), in summarizing ten years of research on the WISC, concluded that any prediction made on the basis of an individual subtest score is little more than a rationalized hunch. Guertin, Rabin, Frank, and Ladd (1962), summarizing five years of research on the WAIS, said: 68 The additional work on scatter, profiles, and patterns has not led us on more solid diagnos tic ground . . . Severe methodological short comings of the investigations prevent the isolation of modal profiles useful for diagnosis CGuertin et al.3 1962). Five years later, in a similar review of the WA.IS literature, the same authors said: Wechsler's paradigms for different diagnostic groups, his Mental Deterioration and Masculinity- Femininity Indices, thus far have produced in terest but little else of practical value (Guertin, Ladd, Frank, and Rabin, 1966). Summary of the Review of the Literature The large number of studies discussed in this partial review in dicates the importance placed upon the use of psychological tests as means for making decisions in educational planning for students who do not make normal academic progress in the regular classroom. The early emphasis in the measurement of intelligence on sensory acuity, grip strength, and memory for dictation gave way to Binet's conception that intelligence was a fundamental quality expressed beha- viorally as ability to judge, understand, and reason. Binet's original objective was to develop a test to identify retardates in the schools. Terman's English language version of the Binet Scale expanded it so that it could also be used for normal and gifted students. Terman's work marked the beginning of systematic individual mental testing in the United States, and the revisions of his test have become the stan dard by which all other intelligence scales are evaluated. David Wechsler, chief clinical psychologist at Bellevue Hospital, 69 developed a series of intelligence scales which were designed for use with mentally ill and mentally retarded. An important innovation of Wechsler's approach was the addition of a Performance Scale to the Verbal Scale which resembled the Binet test. Together the Verbal and Perfoimance Scales provided a Full Scale I.Q. value which represented a gross index of intellectual functioning. A major feature of Wechsler's work, so far mostly unsuccessful, was that of using subtest patterns for purposes of diagnosis and prediction. One of the major issues in the area of mental retardation has been the criteria for classification. Four frames of reference were pre sented: (1) psychometric definitions based on I.Q.'s; (2) legal defi nitions based on the extent to which an individual can manage his own affairs and avoid breaking the law; (3) social competence approach ad vanced by Doll and his followers at the Vineland Training School; and (4) educational definition based upon the individual's ability to per form successfully in various school tasks. Retardation has been found to be most incapacitating during the school attendance years. Mentally retarded individuals have tended to drop out of school short of high school graduation, but once out of school, the mildly retarded have usually been able to get and hold jobs, although they are nearly always at the bottom of the economic scale, are the last to be hired and the first to be dismissed in times of job shortage. The introduction of special classes for the retarded has greatly increased the length of school attendance for the retarded, but on the 70 other hand, the shrinking of job opportunities for the unskilled workers as a result of recent technological change has decreased their chances of employment. The question as to whether educable retardates should be in spe cial classes or in regular classes remains unanswered. Some studies indicate that such retardates achieve better in regular classes while other studies favor special segregated classes. There is a tendency to overlook the differences in educational objectives in the two types of placement. Intelligence has been found to correlate positively with socio economic status, occupational status, parent's socioeconomic and occu pational status, urban living, amount of education, parents' education, intelligence of parents, marital status of parents, regularity of school attendance, living in the North, and being a member of the white race. Intelligence correlates negatively with isolation of residence and number of siblings. A vast body of literature has demonstrated the superior intelli gence test scores of members of the white race as compared to non-whites when measured by instruments composed of items reflecting the white culture. Persistent efforts to develop "culture free" and "culture fair" intelligence tests have not been successful. Furthermore, some researchers have pointed out that validity in predicting success in school learning may be lost in a test which attempts to remove all cul tural influences. A major issue regarding race comparisons in intelli gence has been whether race differences are due to heredity or to 71 environment. Persuasive arguments have been offered by the proponents of both positions, often using the same data. After years of debate a number of researchers in this area came to the conclusion that the question was unanswerable because the crucial variables could not be controlled. More recently, Hupt (1961, 63) has made a strong case for the interaction hypothesis. Verbal intelligence scales have tended to penalize Mexican and Negro children somewhat more than Performance scales, but the Verbal scales are much better predictors of school performance. Mexican and Negro children are over-represented in EMR classes. Non-Spanish-speak ing psychologists have been found to be much more likely to recommend EMR placement for Mexican-American students than Spanish-speaking psy chologists. Some research has suggested that Performance type tests such as the Goodenough Draw-a-Man and the Leiter International Scale be given greater consideration than they are now given in making special class placement recommendations. The "language barrier" is a serious problem for Mexican students and its effects appear to continue as the child, goes up the grades. Virtually all bilingual students appear to be somewhat handicapped by it. One suggestion which has been tried in an effort to reduce the effects of bilingualism is that of inserting an extra year of readiness training between kindergarten and the first grade for Mexican students. Considerable emphasis was placed on the conflict between home values and school expectations which the Mexican and Negro students face. Teachers of minority students have been urged to become acquainted with the value 72 systems of their students' cultural group so that they can more effec tively teach them. The WISC and the WAIS were compared in terms of their standardi zation, administration, scoring, subtest make-up, reliability, and validity. Fisher's research comparing WISC I.Q.'s with those of the Binet and those of the WAIS for institutionalized retardates was pre sented along with his conclusion that the WAIS was invalid for use with retardates. Webb's studies comparing WISC and WAIS I.Q.'s of educable retardates in the public schools were described along with his conclu sion that the WAIS was the more valid test for use with retardates. The subtests of the WISC and the WAIS were described and compared, and the rationale for each subtest was presented along with observa tions of researchers. Several studies of subtest contribution to Full Scale I.Q.'s were presented. Analysis showed that for various levels of retardates the typical subtest pattern was one of non-uniformity in terms of contribution. An attempt to study the relationship between subtest pattern and levels of reading disability produced too many ex ceptions to be very useful for individual appraisal. Another study dealing with subtest contribution of retardates found that Comprehen sion and Similarities were relatively high and Arithmetic and Vocabulary were relatively low among the Verbal subtests, while Picture Completion and Object Assembly were relatively high and Picture Arrangement was relatively low among the performance subtests. CHAPTER III METHOD Selection of the Sample Source The Compton Union High School District was selected as the sample source for a number of reasons. (1) A number of teachers, counselors, and administrators had expressed concern over the frequency with which students had been found ineligible for EMR classes on the basis of I.Q.'s and yet who appeared to be quite unable to profit from regular class- work. Many of them had expressed doubts as to the relevance of the intelligence tests as a predictor of academic success. (2) The Compton District has a large number of each of the three race groups used in the study. (3) Administrators and counselors have long been concerned about the difficult problems met in attempting to provide a really meaningful curriculum for minority race students within the limitations, restrictions, and requirements imposed by a middle class-dominated edu cational structure. These educators had expressed their willingness to cooperate in providing testing space and in releasing students from classes for the required testing time. (4) The investigator had been a member of the District psychological staff for a number of years and it seemed that the data collection and testing could be done with fewer complications here than in another district. Compton is located just southeast of greater Los Angeles within Los Angeles County. It is adjacent to the area that became famous in the 73 74 1965 Watts riots, but Compton itself did not become a part of the riot ing to any significant degree. A few years ago Compton was inhabited almost entirely by members of the white race, but recently members of minority races--primarily Negro and secondarily Mexican--have been moving in and the whites have been moving out. The change in racial composition of the city has been greatest on the west side and decreas ing toward the east. Today there are almost no white students attending the schools in the western and central sectors. The white majority in the one senior high school (Dominguez) and the two junior high schools (Roosevelt and Whaley) on the eastern edge of Compton is gradually being replaced by Negro and Mexican students. In some cases children have been sent here to live with relatives by parents who remained behind in the South. The great difficulty encountered by a school district which must absorb large numbers of such students into its program can be ap preciated in the light of studies which have found average I.Q.'s of 69 for rural southern Negro children (Young and Bright, 1954). The Compton Union High School District is composed of seven junior high schools and three senior high schools. These schools range in en rollment from 1040 to 2780. The total enrollment for the District was a little over 16,000 as of February, 1967. No data were available concerning the racial composition of the District or of the individual schools. Many students, parents, teachers, administrators, and clerical employees who are members of minority races resent attention being drawn to racial differences. It has been the policy in the Compton Schools (as well as other similar districts with 75 which the investigator is acquainted) to omit racial categories in their statistics. Table 4 shows the average daily attendance for the period from January 30, 1967 to February 24, 1967. TABLE 4 COMPTON UNION HIGH SCHOOL DISTRICT DISTRIBUTION OF ATTENDANCE (AS OF FEBRUARY, 1967) AVERAGE DAILY ATTENDANCE School Regular Classes EMR Classes Total Percent EMR (Junior High) ■ Bunche 854 150 1004 14.9 Enterprise 1101 101 1202 8.4 Roosevelt 1241 94 1335 7.0 Vanguard 1198 90 1288 7.0 Walton 968 135 1103 12.2 Whaley 1147 50 1197 4.2 Willowbrook 879 141 1020 13.8 Total J.H.S. 7388 761 8149 9.3 (Senior High) Centennial 1794 168 1962 8.6 Compton 2340 145 2540 5.7 Dominguez 1899 54 1953 2.8 Total S.H.S. 6033 367 6455 5.7 Total Junior and Senior High 13421 1128 14604 7.7 In addition to the 14,604 total attendance shown on the table there were 55 students attending classes for the trainable mentally- retarded (TMR) and 404 units of attendance credited for students in the adult and continuation programs. Population Number One Purpose Population number one was used in an effort to study the follow ing questions: 1. Will below-average intelligence subjects obtain different | Verbal, Performance, and Full Scale I.Q.'s on the WISC and ; the WAIS? 2. Will there be interaction by race and sex groups? j ; i 3. To what extent will any difference between the WISC and the WAIS Full Scale I.Q.'s result in conflicting recom- j ; i I mendations? i j | i 4. Which subtests account for any differences in Full Scale ' i i. q .? I I I Description j ; ! Population number one was comprised of Anglo-American, Mexican- j j ; American, and Negro-American boys and girls who were within three months I of the sixteenth birthday and who had scored below 90 on a recent group ! jor individual intelligence test. 77 Selection of the Sample Nearly all of the students in the Compton Union High School Dis trict have fairly recent California Test of Mental Maturity scores available in their cumulative record folders which could be used to identify subjects suitable for the sample. EMR students were not or dinarily given group intelligence tests and their individual intelli gence test scores were considered as qualification for the study. The term "recent" (test) w^s defined as not more than three years old at the time of the sample selection. Students enrolled in special classes for the trainable retarded or in special classes for the blind, par tially- seeing, deaf, hard-of-hearing, or physically or neurologically i ' ! ! handicapped were not included. j j By District policy a student's race was not written on any of his j i records. Therefore, the investigator inspected the cum folders of all j ninth, tenth, and eleventh grade students in the ten schools in the Dis-i jtrict and wrote on index cards the name, birthdate, sex, grade, school, land most recent CIMM scores for each student who satisfied the pre- | jviously described criteria. In most cases the race of the student could ! | : be tentatively identified by the collection of pictures present in the I i I icurn folders or by the first and last names of the student, parents, and! : i siblings. Inmost of the doubtful cases the student's grade counselor | I i |was able to provide the correct information. Several Spanish-speaking, ! s ■ i : Mexican-born Negroes were not included. Orientals and students with j mixed Mexican-Negro parentage were also not used. The index cards were arranged chronologically by date into six 78 race-sex groups: Anglo male, Anglo female, Mexican male, Mexican female, Negro male, and Negro female. Cards for subjects bom between September 1, 1950, and January 31, 1951, were used to foim the initial reserve pool from which ten cards from each race-sex group were randomly selec ted for the testing list. Since the testing was to begin in November of 1966, all of these subjects were within three months of the six teenth birthday. When a subject’s card was removed from the testing list as a result of his being tested or found to have moved out of the District, a replacement was randomly selected from the reserve pool. On January 1, 1967, cards for students bom in March, 1951, were added, and so on. As time passed, the cards were eliminated from the reserve pool for those students who had passed the point of sixteen years three j i I months of age. When twenty subjects in any race-sex classification had | i ; I been tested, the remaining cards for that group were withdrawn from the ! testing list and the subject pool. Testing continued until twenty sub- ! : ' ! i jects had been tested from each group, a total of one hundred twenty in! i i all. : The Testing of Sample Number One Test Examiner ! Originally it was hoped that the race-of-examiner effect could jbe controlled by dividing the testing equally among examiners of the ; jAnglo, Mexican, and Negro races; however, the responsibilities of the ' other Compton District psychologists prevented their participation and j no other psychologists or psychometrists could be obtained. In addition, several members of the administrative and counseling staff at the 79 schools expressed reluctance to permit anyone not directly employed by the school district to test their students, especially since the test ing would be done in their offices and the students would be released from classes for the time required. However, some of the testing was done by a member of the district teaching staff who was doing field work jin school psychology under the supervision of the investigator. So all |of the testing and scoring of the tests of the subjects in sample one was done by, or under the supervision of, the investigator. Place of Testing Although most of the testing was done during school hours in con ference rooms or counseling offices on the school premises, some of the (students voluntarily remained after school, came in the evening, or on Saturdays, Sundays, or on school holidays. This was true in all cases i f where the testing was done by the psychologist-in-training since he was teaching a full load of classes. Testing outside of the regular school j hours did not increase the sample influence of volunteer subjects be- I J cause non-volunteers were tested during school hours. j Testing Procedure The routine individual testing procedure was followed. The coun- jselor summoned the student from class, brought him to the place of testing, and introduced him to the examiner. The examiner explained i ithat the purpose of the testing was to evaluate some of the standard ! instruments and that, while the student was free to decline to parti- j ! i cipate, his taking the tests would be of value to the District. It was j further pointed out that the student would benefit by having individual 80 test records in his cum folder when discussing future educational and vocational plans with his counselor. None of the one hundred twenty subjects declined. Due to the problems that would be involved in scheduling two test sessions for each subject and the attendant likely loss of some subjects as a result of their moving out of the district, both tests were administered on the same day. In order to control test effects the numbers one through eleven were assigned to the WISC subtests (omitting Mazes) and the numbers twelve through twenty-two were assigned to the WAIS subtests, and the twenty-two subtests were administered to each subject in random order. This procedure was chosen because of statements in the test manuals. The WISC manual states "variations in order are permitted" (Wechsler, 1949, 19). The WAIS recommended that the regular order be followed but provided "in some rare instances it may be preferable to vary the order of presentation" (Wechsler, 1955, 27). The usual procedure which was followed was to take a break be tween the eleventh and twelfth subtests, but there were some unavoidable interruptions such as the lunch or recess periods. Test administration to two subjects was interrupted for about ten minutes by fire drill, but fortunately in both cases the interruption did not occur during a timed subtest. Another question which nay legitimately be raised concerning the procedures in this study would be that of the effects of fatigue as a result of taking both tests on the same day. Some subtests might be influenced more than others by fatigue. For example, a timed test such 81 as Coding might be more vulnerable to the effects of fatigue than an untimed test like Information or Comprehension. Carrying this line of reasoning a little further, the WISC Coding is timed for two full min utes and thus may suffer more fatigue effects than the WAIS version (Digit Symbol) which is sustained for only one and one-half minutes. The investigator attempted to reduce the effects of fatigue by observing the subjects carefully in an effort to detect signs of weariness and by asking each subject to notify him if he felt tired and wanted to take a break. In addition, as previously mentioned, in all cases there was a break of at least fifteen minutes in the middle of the testing. The Data from Sample One The data from sample number one consisted of the race, sex, grade ! ■ I :in school, type of placement, C1MM I.Q.'s (Language, Non-language, and j Total), and the WISC and WAIS I.Q.'s for the three scales (Verbal, Per- j formance, and Full Scale) and the scale scores for each of the twenty- ;two subtests. All of the data except those pertaining to the WISC and j i s iWAIS were obtained from inspection of cum records and conferences with j :counselors. The WISC and WAIS data were obtained by administering the : |two tests to each of the one hundred twenty subjects and scoring the i protocols in accordance with the instructions in the mr Is. ; . | ' Preparation of the Data for Analysis i ! The data from the subject information cards and test protocols ! i I . iwere transferred to work sheets and then key-punched into IBM cards. The coded information punched into cards for the sample from population, 82 one was as follows: Race Sex Identification Number Type of Placement (Regular or EMR Classes) Reading Achievement Level CIM4 Language I.Q. CTMM Non-language I.Q. CIMM Total I.Q. WISC Information WISC Comprehension WISC Arithmetic WISC Similarities WISC Digit Span WISC Vocabulary WISC Verbal I.Q. WISC Picture Completion WISC Picture Arrangement WISC Block Design WISC Object Assembly WISC Coding WISC Performance I.Q. WAIS Information WAIS Comprehension WAIS Arithmetic WAIS Similarities WAIS Digit Span WAIS Vocabulary WAIS Verbal I.Q. WAIS Digit Sumbol WAIS Picture Completion WAIS Block Design WAIS Picture Arrangement WAIS Object Assembly WAIS Performance I.Q. WAIS Full Scale I.Q. Testing Hypothesis Number One i : The questions to be answered with the data obtained from the j | sample from population number one were: (1) Will below-average intellij | gence individuals obtain higher WAIS than WISC Verbal, Performance, andj Full Scale I.Q.'s when taking both tests? (2) Will the differences ; between the two tests be greater for the members of one racial group 83 than for another? (3) Will the difference between the two tests be greater for one sex than for the other? (4) Will there be interactions between race and sex? On the basis of previous studies (Fisher, 1961, and Webb, 1963, 1964) and the professional experience of the investigator and his col leagues in using the two tests in the public schools, the predictions were made that: (1) WAIS will surpass WISC on Verbal, Performance, and Full Scale I.Q.’s for the sample as a whole and for all race, sex, and race-sex sub-groups. (One tail test. p=.05) (3) There will be no dif ference between the WAIS minus the WISC I.Q. differences for one sex group as compared to the other. (Two tail test. p=.05) (4) There will be no interaction between race and sex. (Two tail test. p=.05) ; ! i i To test these predictions an analysis of variance technique was iemployed. The Practical Effects on Placement Recommendations of WISC-WAIS Test Differences , In practice, I.Q. has often been the primary if not the sole de- I terminer of EMR class placement. The rationale for this dependence on ; i I 'I.Q. seems to be that an individual who scores below a certain point i : | (e.g. 80) on an I.Q. test is considered to be incapable of profiting : ! ' ; ;from ’ ’ regular class instruction” whereas another individual who scores ; j | i above that point, even though he may have severe academic deficiencies,; I I j does not qualify for EMR placement. Basing predictions of academic j j i : success or failure on the I.Q. in effect implies that the I.Q. can be ; measured with some degree of accuracy. Recent findings that two of 84 the most widely used and respected I.Q. tests (the WISC and the WAIS) produce significantly different values for the same subjects when tested several years apart raised the question for which the investigator is attempting to find at least a tentative answer in this portion of the present study. In practice, what proportion of a . group of below-average intelligence subjects would have conflicting placement recommendations by the two tests? To answer this question the one hundred twenty subjects in the sample were categorized in terms of their Full Scale WISC's and WAIS's: (1) those with both I.Q.'s above 79 or both I.Q.'s below 80 (thus both tests having the same recommended class placement); (2) those with WISC Full Scale I.Q. below 80 and WAIS Full Scale I.Q. above 79 (WISC recom- ! Imending EMR and WAIS recommending regular class placement); and (3) those with WISC above 79 and WAIS below 80 (WISC recommending regular I classes and WAIS recommending EMR class placement). These data were ;then converted to percentages. : i Subtest Contributions to WAIS-WISC Differences j i ; Which subtests accounted for the differences between the WISC and | WAIS Full Scale I.Q.'s? To answer this question the following procedure • ! jwas followed: 1. The WISC and WAIS subtest means were computed. In order to make 1 . i comparisons of WISC and WAIS scale score means it was necessary to ; j ; convert the WAIS raw scores into scale scores having a mean of ten and a standard deviation of three. This was done using the special conver- | sion table provided for this purpose in the WAIS manual. 85 2. Subtest differences were obtained by subtracting each WISC sub- test mean from the equivalent WAIS., subtest mean. 3. The subtest mean differences were summed. 4. Each subtest difference value was divided by the sum of the sub- test differences to deteimine the percent of the total differences which was contributed by each subtest pair. 5. The eleven subtests were ranked according to the size of their contributions to the WAIS-minus-WISC Full Scale I.Q. differences. 6. The data were examined in greater detail to see whether race* sex, and race-sex groups tended to follow the general pattern of subtest difj ference contributions of the sample as a whole. | < i Population Number Two Description of Population Number Two ; i Population number two is comprised of students who had been re ferred for testing to the District psychological examiners as suspected) , ' l imental retardates and who had been administered both an individual j I i I ! i reading achievement test and either the WISC or the WAIS. The function j j of population number two was to provide a sample from which data could ! ;be obtained to test the hypothesis that individual reading tests could I ; j i ! I be used more effectively than individual I.Q. tests to predict success ; : i :in regular classes for suspected mental retardates. j j ’ I ! The Sample from Population Number two j i | ! The sample consisted of all the students in population number two‘ for whom a determination of ’ ’ graduate" or "dropout" in regular or EMR " " ' ’ ’ ’ ~~ 8 6 ” | classes could be made for the years 1965, 1966, and through July of 1967. The investigator screened approximately two thousand case studies in the District clinical files and wrote on index cards the name, sex, birthdate, test age, school, grade level at the time of testing, name of I.Q. test, Verbal, Performance, and Full Scale I.Q.'s, name of reading test, reading achievement level, and type of placement (regular or EMR classes) for all students whose files contained all of the above information and who were bom in the year 1947, 1948, or 1949. These birth years were selected because the students bom during this period j would have had a chance to graduate or to drop out by the time the J data were gathered in July of 1967. The cards were then sorted accord-j ; j ! ing to school attended and traced by inspecting the permanent records i ;and other District records to determine whether each student had (1) graduated, (2) had dropped out of school with no subsequent request] I for transfer of records to indicate that they were attending another | j ’ i J school, (3) had transferred out of the District, and (4) was still at- j ! j 1 tending school as a result of not yet meeting graduation requirements. j ] The first two categories were the ones of interest since category one ; ! meant "success" and category two meant "failure." Students in cate- igories three and four were eliminated from the sample since it could j ! not at that time be determined whether they were, or would be, successes i ! or failures. 87 Gathering the Data The coded information punched into cards for the sample from population number two was: Race Sex Identification Number Chronological Age Test (WISC or WAIS) Verbal I.Q. Performance I.Q. Full Scale I.Q. Reading Achievement Level Actual Grade Placement at the Time of Testing Reading Index Type Placement (Regular or EMR Classes) Grade at Disposition Type Placement at Disposition (Regular or EMR Classes) Disposition (Graduate or Non-Graduate) The I.Q. tests to be used for comparison were the WISC and the WAIS, and the reading achievement tests were the Wide Range Achievement Tests (Jastak, 1946) and the Gilmore Oral Reading Test (Gilmore, 1951), an individual reading test. The Gilmore provides an Accuracy score and a Comprehension score while the Wide Range has a single achievement grade level score. In order to equate the Gilmore and the Wide Range, the two Gilmore values were averaged. In a few cases where both the Gilmore and the Wide Range had been given in the same test battery, the two Gilmore scores were averaged and then that score was averaged in turn with the Wide Range so that each subject would be represented by a single reading achievement score. Of the approximately 2000 cases, 753 were found to have birth dates within the 1947 to 1949 range. Of these, 183 graduated, 197 dropped out of school with no request for transfer of records to 88 another school district, 208 were considered to have transferred since transcripts had been requested by other school districts, and 89 were still attending classes within the district. For the remaining 76 cases, no records indicating disposition could be found. In going through the permanent records at the senior high schools, the investi gator noticed that a number of students had file numbers but the records were not in the assigned file books. In addition, many cards were filed in the wrong book, so that it must be assumed that many of the missing records are somewhere in the school files. There were a number of cards for students who were purported to have moved from the ninth |grade in one of the junior high schools up to the tenth grade in a senior high school in the district, but the senior high school had no ! ' ! I record of their ever attending. ! The effect that these factors would have on the data in this study would be to increase spuriously the proportion of graduate as opposed to non-graduates, since all of the subject cards were first i | checked against the official graduation lists in the senior high schools ; I i I ! in an effort to reduce the amount of tracing that would have to be done; individually through the files. In other words, a . graduate would be likely to be counted in the data even though his permanent record file i was lost or misplaced, but a non-graduate whose file was lost or mis- | placed would be dropped from the sample as an "unknown" as far as ' i I ! disposition is concerned. However, there is no reason to believe that * I this somewhat inflated "success" category will systematically bias the ’ data in favor of one test predictor over the other. It does mean that 89 that the data from this study could not accurately be used to indicate the proportion of students who graduate from this particular population. Gathering the Data for Sample Two To compare the reading tests and the I.Q. tests as predictors of success in regular classes, certain criteria were established. "Suc cess" was defined as graduation from the twelfth grade and "failure" was defined as dropping out of school short of graduation. The cutoff point for prediction by the I.Q. tests was the one used in practice in the schools. An I.Q. of 80 or above would predict "success" in regular classes and an I.Q. below 80 would be considered to predict "failure" in regular classes. A problem was encountered in comparing reading achievement test scores to I.Q.'s. The I.Q. takes into account the subject's age, where as reading achievement does not. In terms of retardation, a seventh grader reading at fifth grade level would not be as handicapped as an eleventh grader reading at fifth grade level. So comparing grade level to I.Q. without regard to grade placement would not be meaningful. Likewise, using the amount of retardation would not be feasible, be cause a seventh grader retarded by two years in reading ability would be more handicapped than an eleventh grader with the same number of years of retardation. Therefore each student's reading achievement score was converted into an index to represent the ratio between his actual grade placement (the denominator of the formula) and his read- achievement level (the numerator). A subject reading at grade 90 level would have an index of 1.00. Actual grade placement was used in the formula in preference to chronological age because actual grade placement more nearly represents the level of academic challenge the student probably encounters in regular classes. Testing Hypothesis Number Two i Hypothesis number two suggests that reading achievement tests can be used more effectively than I.Q. tests to predict success in regular classes for suspected mental retardates. To test this hypothesis the Biserial correlation technique was found to be appropriate because the data meets the following necessary assumptions: (1) continuity in the dichotomized trait (2) normality of distribution underlying the dichotomy (3) a large N (4) a split that is not too extreme j The formula used in computing the Biserial correlations is: ; r - M -M X p q bis p q S.D. u in which M = the Mean of the graduates M = the Mean of the dropouts S.D. = the standard deviation of the entire group p = the proportion of the entire group who are in the graduate category q = the proportion of the entire group who are in the dropout category u = the height of the normal curve dividing the two parts, p and q (Garrett, 1961, 378) 91 In order to determine whether the difference between the reading index and I.Q. correlations with graduation were significant, the cor relation values were converted into Fisher's function and the signifi cance between z values was determined using the formula fv = tfzj - Z2 -\l_l_ - 3 + _1_ - 3 N1 N2 (Standard Error of the difference between z coefficients) (Garrett, | i 1961, 241) Because the biserial correlation coefficients have been found to be spuriously high, the values required for significance in the tables were increased by 25 percent as a precaution. Practical Effects of Using Cutoff Points in Prediction The investigator wished to determine the practical effects of the j * | use of I.Q. and reading indices as predictors of academic success for suspected mental retardates. An arbitrary cutoff point of .50 was se lected for the purpose of comparing the reading test to the I.Q. tests Ssince no precedent value could be found. In other words, if an indi- j i I jvidual's reading achievement level were as much as half his actual grade j : (placement level, then the reading index would be considered as predict- j ling "success" in regular classes for that individual; and if his reading ! achievement level were below that point, the reading index would be con-; : ’ I sidered as predicting "failure." It was expected that some value other j ithan .50 would prove to have a higher degree of efficiency but for the ; jpurpose of comparison in the study, the value of .50 was used. j ' i The I.Q.'s were converted into "predict success" for each subject I j with I.Q. of 80 or above and "predict failure" for those with I.Q. below 80. By comparing this prediction to the actual outcome for each subject, computing the number of cases and the number of correct pre dictions, and dividing the number of correct predictions by the number of cases, the per cent of accurate prediction by I.Q. was obtained. Separate treatment was given to the WISC and the WAIS as well as for the Verbal, Performance, and Full Scale I.Q.’s i In the same manner reading indices were converted into "predict success” for each student with index of .50 or better and "predict failure" for those with index below .50. By comparing this prediction to the actual outcome for each subject and dividing the number of cor rect predictions by the number of cases, the per cent of accurate pre- : diction by the reading index was obtained. The significance of the ! : i difference in accuracy of prediction between I.Q. and reading index j was determined by the following procedure: (1) The standard error of a percentage was computed for the per cent of accuracy value for the Full Scale I.Q. and I i for the reading index by the formula j | | i ( F P = \ | P Q ! \ N ! where P = the percent of accuracy I Q = 100 - P j N = the number of cases j | (Guilford, 1956, 175) j ; ■ j (2) The standard error of a difference between percentages was j computed by the formula i 93 where ^ = the standard error of the first percentage r = the standard error of the second percentage OP2 (Guilford, 1956, 191 (3) The z value was obtained by the formula <fd P where = the larger percent of accuracy P2 = the smaller percent of accuracy ^ = the standard error of the difference between 0 dp percentages (4) The z value was applied to the appropriate table (Guilford, 1956, 533) to determine the level of significance of the difference between the accuracies of the two instruments. CHAPTER IV FINDINGS Hypothesis One I Hypothesis one is concerned with an examination of the compara bility of the WISC and the MIS I.Q.'s for below-average intelligence subjects. In this section the dependent variable is the obtained I.Q. for each subject on each of the two tests. The values for the one hundred twenty subjects in the sample are included as Exhibits 1 through 4 in Appendix A. Tables 5, 6, and 7 summarize these data presenting [means, standard deviations, and ranges on the Verbal, Performance, and i j Full Scales respectively. ■ TABLE 5 VERBAL I.Q.'s FOR SAMPLE ONE Mean WISC S.D. Range Mean WAIS S.D. Range All Subjects (N=120) 80.11 10.76 58-108 87.83 9.49 70-107 Anglo (n=40) 82.55 12.52 58-108 90.47 10.96 70-107 Mexican (n=40) 77.65 8.94 60-99 83.92 8.05 72-102 Negro (n=40) 80.15 10.20 60-108 89.10 8.09 72-104 Male (n=60) 81.41 11.45 60-108 88.43 9.62 71-107 Female (n=60) 78.81 9.94 58-104 87.23 9.40 70-107 94 95 According to Table 5 the mean Verbal I.Q. for all subjects is nearly eight points higher on the WAIS than on the WISC. On the other hand, the WISC variability is greater than that of the WAIS. The above observations concerning the sample as a whole apply to each of the three racial groups without exception. In terms of inter racial comparisons, the Anglo mean Verbal I.Q. was higher than the Negro which in turn suipassed the Mexican. This was true for both the WISC and the WAIS, but the Mexican disadvantage was somewhat greater on the WAIS. The Anglo subjects had the greatest variability, followed by the Negroes and then the Mexicans. Both sexes had higher WAIS than WISC mean Verbal I.Q. 's and the WISC variation was larger than that of the WAIS also. Males surpassed j females in mean Verbal I.Q. on both tests. Males had slightly larger |standard deviations than females as well as the greater range on the ! iWISC. However, females had a slightly larger range on the WAIS Verbal. : TABLE 6 PERFORMANCE I.Q.'s FOR SAMPLE ONE Mean WISC S.D. Range Mean WAIS S.D. Range All Subjects (N=120) 88.51 13.96 51-125 90.95 9.98 70-114 Anglo (n=40) 91.72 12.71 57-113 93.12 9.76 78-114 Mexican (n=40) 88.25 14.23 55-125 90.80 9.15 74-108 Negro (n=40) 85.57 14.53 51-118 88.25 10.77 70-107 Male (n=60) 90.85 15.10 55-125 92.01 10.35 70-108 Female (n=60) 86.18 12.40 51-118 89.88 9.56 72-114 96 According to Table 6 the mean WAIS Performance I.Q. for all sub jects is slightly (2.44 I.Q. points) larger than that of the WISC. As in the case of the Verbal scale, the WISC variation is much larger than that of the WAIS. The WISC standard deviation is nearly four I.Q. points larger than that of the WAIS, while the WISC range extends nineteen I.Q. points below and eleven points above that of the WAIS. ! ! For all three race groups the WAIS mean Performance I.Q. is larger I than the WISC, while the WISC standard deviations and ranges are larger than those of the WAIS. In terms of racial comparison, the Anglos have the highest performance I.Q.'s on both tests, while the Mexicans are second on the WISC and the Negroes are second on the WAIS. The Mexican and Negro standard deviations on the WISC Performance are nearly as large as that of the norm population in spite of the restricted range of this sample. The standard deviation on the WISC Performance for the j : I Anglos is somewhat smaller. The WAIS Performance standard deviations ' are very much smaller than those of the WISC and the racial ranking is i jNegro largest, then Anglo, and Mexican smallest. The WISC ranges were | ! ’ ! iconsiderably larger than the WAIS ranges for all three races. On the j i • • j :WISC range, the Anglo group was considerably smaller than the Mexican | i i i I land Negro group, but there was little difference among the groups on ■the WAIS. | ; | I Males surpassed females by about five points on the WISC Perfor- j : i 1 j jmance I.Q. and about two points on the WAIS. Males had considerably I j j I greater variance than females with males having a standard deviation ! equal to that of the noim population (15). Both groups had much 97 smaller standard deviations on WAIS Performance with the males slightly higher than females. TABLE 7 FULL SCALE I.Q.'s FOR SAMPLE ONE Mean WISC S.D. Range Mean WAIS S.D. Range All Subjects (N=120) 82.52 11.17 51-109 88.80 8.61 71-107 Thiglo (n=40) 85.55 11.51 58-107 91.50 9.28 73-107 Mexican (n=40) 81.02 10.61 58-109 86.35 7.26 73-105 Negro (n=40) 81.00 11.01 51-101 88.58 8.59 71-103 Male (n=60) 84.46 11.79 58-109 89.95 8.72 73-107 Female (n=60) 80.58 10.24 51-101 86.67 8.42 71-107 According to Table 7 the mean Full Scale I.Q. for all subjects is 6.28 points higher for the WAIS than for the WISC. The variance is jlarger for the WISC than for the WAIS with noticeable differences in j ; jboth the standard deviations and the ranges. The largest difference.-in- ! ' ! |ranges is at the bottom of the distribution where the WISC range extends 1 • twenty points lower than that of the WAIS whereas at the other end it i : extends only two points above the WAIS, For all three race groups the WAIS Full Scale I.Q. surpasses that! j of the WISC while the standard deviations and ranges are greater for i ! ! : the WISC than the WAIS. The Anglo group is about four and one-half J • ' ! I.Q. points above the Mexicans and Negroes on the WISC, whereas they surpass the Negroes by three points and the Mexicans by five points on 98 the WAIS. There is little difference in variation among the racial groups for the Full Scale I.Q. Apparently the much greater differences in the Verbal and Performance Scales tended to cancel out. Males surpassed females by about four I.Q. points on the Full Scale WISC and by about three I.Q. points on the Full Scale WAIS. There was almost no sex difference in range and little difference in standard deviation. Hypothesis number one was tested by analysis of variance based on the technique discussed by Hartley in Mathematical Method for Digital Computing (Ralston, 1962, Chapter 20). Table 8 summarizes the signi ficant main effects and interactions abstracted from the complete j {analysis of variance summary tables included as Exhibits 1 through 3 in Appendix B. TABLE 8 SUMMARY OF SIGNIFICANT ANALYSIS OF VARIANCE MAIN EFFECTS AND INTERACTIONS PERTAINING TO THE COMPARABILITY OF THE WISC AND WAIS I Source of Variation Verbal Performance Full Scale {Race Overall Differences Between Tests 3.685* 236.969*** 2.388* 8.172** 3.236* 112.998*** i (Abstracted from Exhibits 1 through 3 in Appendix B.) ;* Significant at .05 level but not at .01 level ;** Significant at .01 level but not at .001 level '*** Significant at .001 level 99 According to Table 8 the overall differences between the WISC and WAIS on the Verbal scales and on the Full Scales are significant beyond the .001 level. Differences between the two tests on the Performance scales is significant beyond the .01 level. Racial differences between the two tests were significant beyond the .05 level but not the .01 level for all three scales. There were no significant sex differences. Table 9 summarizes the WAIS minus WISC differences for all groups on the three scales. ] TABLE 9 I WAIS MINUS WISC DIFFERENCES I I i i ■ ' " ~ ~ • • — . . . . _ ! j ! Verbal Performance Full Scale ! I. . i ■ — ■ ■ ■ ■■■ ■ i | ] All Subjects 7.72 2.44 6.28 ! : ' \ Anglo 7.92 1.40 5.95 Mexican 6.27 2.55 5.33 i Negro___________________8.95_______________5.35_________7.58______; i ; | Male 7.02 1.16 5.49 j i Female__________________8.41______________ 3.70_________6.09______; I i According to Table 9 for all subjects the WAIS Verbal I.Q.'s ex- J j ' ! ceeded those of the WISC by nearly eight points. The WAIS over WISC ' i { ; Performance Scale advantage was much smaller (2.44 points), while the ! I j J Full Scale difference in favor of the WAIS was slightly less than six | t I i ! I.Q. points. | All of the race groups followed the WAIS surpassing WISC pattern j for all three scales. The Negroes' mean Verbal MIS was almost nine 100 I.Q. points higher than the WISC, while the Anglos' was almost eight and the Mexican' slightly more than six. The Negroes also had the greatest disparity on the Performance (3.32 points) followed by the Mexicans with 2.55 and the Anglos with very little difference. The order of magnitude of difference on the Full Scale I.Q. was Negro, Anglo, and Mexican, j - In terms of sex differences, the females had a somewhat larger WAIS minus WISC difference on all three scales. Matrices of the Pearson product moment correlation coefficients between the scales of the two tests for all subjects, race groups, sex groups, and race-sex groups are included as Exhibits 1 through 4 in Appendix C. Table 10 presents a summary of these values extracted from the exhibits. i TABLE 10 ! PRODUCT-MOMENT CORRELATIONS BETWEEN WISC AND WAIS FOR SAMPLE ONE Verbal Perfoimance Full Scale All Subjects (N=120) .76 .72 .80 'Anglo (n=40) .77 .80 ■ .83 Mexican (n=40) .55 .67 .76 Negro (n=40) .89 .70 .80 Male (n=60) .67 .71 .78 Female (n=60) .86 .74 .82 Anglo Male (n=20) .66 .71 .74 jAnglo Female (n=20) .87 .83 .88 Mexican Male (n=20) .50 .70 .79 Mexican Female (n=20) .69 .57 .52 Negro Male (n=20) .87 .66 .75 Negro Female (n=20) .91 .75 .85 101 . > These correlation values appear to be comparable with those ob tained in previous studies of Webb and others. For all subjects the Full Scale correlations were higher than those of Verbal and Performance Scales. By race the Negro WISC-WAIS correlations were the highest while those of the Mexicans were the lowest. By sex the WISC-WAIS cor relations of the females surpassed those of the males. Carrying the j analysis further, WISC-WAIS correlation on the Verbal Scales is high for Negro females, Negro males, and Anglo females, and low for Mexican { I males. Correlations on the Performance Scale are somewhat smaller than on the Verbal. Only the Anglo female and possibly Negro female could be considered fairly high, while the Mexican female would be considered i I fairly low. On Full Scale correlations Anglo female and Negro female I | j are high, while Mexican female correlation is fairly low. j ; In view of the highly significant differences between the WISC ; and the WAIS on the Verbal and Full Scales and the significant (.05) 1 differences between the two tests on the Performance Scale, it was de- : | ■ i I cided to examine the data to see what practical effect might occur in j ! terms of placement as a result of these differences. In accordance ! i ' ; i 1 j with what appears to be the most frequently used cutoff point, all ! | I j Verbal, Performance, and Full Scale I.Q.'s 80 and above were coded , 1 I i "regular class placement" and all those below 80 were coded "EMR" j j I | placement." Table 11 presents these data. J 102 TABLE 11 WISC-WAIS DIFFERENCES IN PLACEMENT RECOMMENDATION VERBAL WISC:REG WISCrEMR SAME WAIS:EMR WAIS:REG PERFORMANCE WISC:REG WISC:EMR SAME WAIS:EMR WAIS:REG FULL SCALE WISC:REG WISC:EMk SAME WAIS:EMR WAIS:RE AM* 13 ! af 15 MM 13 MF 10 ! NM 10 NF 13 0 0 1 0 0 0 7 20 0 0 19 5 15 1 4 15 6 18 0 2 15 10 14 0 6 13 10 15 1 4 14 7 16 0 4 13 0 0 0 0 0 0 1 5 5 7 6 7 Sum 74 1 45 j j % 61.7 0.8 37.5 98 2 20 81.7 1.7 16.67 89 74.17 0 0 31 25.83 AM=Anglo Male AF=Anglo Female MM=Mexican Male MF=Mexican Female NM=Negro Male NF=Negro Female According to Table 11 in 74 (61.7%) of the one hundred twenty cases ■ i the WISC and WAIS Verbal Scales would make the same class placement ; i recommendation, whereas in one case (a Mexican male) the WISC would re- j jcommend regular classes and the WAIS would recommend EMR class placementj. Iln 45 cases (37.5%) the WISC Verbal would recommend EMR while the WAIS i would recommend regular class placement. All race-sex groups had five j i ! jor more cases of this disagreement in placement recommendation, but the I Mexican females and Negro males had the most with ten cases each. There 1 i !was considerably less WISC-WAIS placement disagreement on the Performance iScales. In 98 cases the two tests made identical recommendations, in ( two cases the WISC recommended regular classes while the WAIS recommen ded EMR. In twenty cases (16.7%) the WISC recommended EMR while the 103 WAIS reconmended regular class placement. The Full Scale I.Q. is the primary one considered in the placement recommendations. In the present study in 89 cases the WISC and WAIS Full Scale recommendations were identical and in no case did the WISC recommend regular classes and the WAIS EMR classes. However, in 31 cases (25.8%) the WISC Full Scale I.Q. would resillt in a recommeridation for EMR classes while the WAIS Full Scale I.Q. would indicate regular classes. This conflicting re commendation occurred only once in the Anglo male group, hut all of the other race-sex groups had five, six, or seven conflicting recommenda tions of the twenty subjects in each group. In order to analyze the data in greater detail in an effort to determine which subtests accounted for the large differences between ithe WISC and WAIS I.Q. ’s, Table 12 was constructed showing the mean ; i jsubtest scale scores for the WISC and Table 13 was constructed showing ! i * ! these values for the WAIS. j I According to Table 12 the mean subtest scale score on the WISC | iCoding was actually a little above the mean of the norm population. In j I ' j jail of the other subtests the sample as a whole performed well below j j I 'average. (A scale score of ten is average.) Aside from the Coding, | ; ' ! Ithe best performance on the WISC subtests was on Object Assembly and I Picture Couplet ion. The sample as a whole performed poorest on Vocabu- j Ilary (6.05), Arithmetic (6.20), and Information (6.27). I j ; j The Anglo group was well above the population mean on the WISC ! ! ! jCoding (10.6) and almost average (9.2) on Object Assembly. Picture j Completion (8.67) was also within the normal range. Their poorest per- j 104 formance was on Information (6.70), Arithmetic (6.80), and Vocabulary (6.97). The Mexicans were average (9.98) in Coding and in the normal range (9.12) in Object Assembly. Their greatest handicap was in Vo cabulary (5.22) and in Arithmetic (5.80), but the Information (6.00) anc Comprehension (6.58) were also poor. The Negroes were average (10.0) on the WISC Coding and in the normal range on Digit Span (8.10), Object Assembly (8.12), and Picture Completion (8.02). Their poorest showing was on Vocabulary (5.95) but they were also low on Arithmetic (6.02), JInformation (6.12), and Block Design (6.50). j TABLE 12 ! WISC MEAN SUBTEST SCALE SCORES 1 . — 11 1 INFO COMP ARI SIMI VOC DISP PICC PICA BLKD OBJA CODG All Subjects 6.27 7.08 6.20 7.49 6.05 7.50 8.22 7.25 7.32 8.82 10.20 Anglo 6.70 7.30 6.80 7.55 6.97 7.45 8.67 7.62 7.88 9.22 10.62 Mexican 6.00 6.58 5.80 7.62 5.22 7.45 7.97 7.10 7.60 9.12 9.98 Negro 6.12 7.38 6.02 7.30 5.95 8.10 8.02 7.02 6.50 8.12 10.00 Male 6.60 7.57 6.20 7.52 6.32 7.48 8.98 7.63 8.02 9.53 9.42 iFemale 5.95 6.60 6.21 7.47 5.78 7.52 7.47 6.87 6.63 8.12 10.98 'Anglo Male 6.90 8.45 6.90 7.95 7.90 7.60 9.85 8.30 8.85 9.95 10.30 Anglo Female 6.50 6.15 6.70 7.15 6.05 7.30 7.50 6.95 6.90 8.50 10.95 Mexican Male 6.65 6.70 6.35 7.40 5.35 7.15 8.35 7.20 8.55 9.50 9.55 Mex. Female 5.35 6.45 5.25 7.85 5.10 6.75 7.60 7.00 6.65 8.75 10.40 iNegro Male 6.25 7.55 5.35 7.20 5.70 7.70 8.75 7.40 6.65 9.15 8.40 l [Negro Female 6.00 7.20 6.60 7.40 6.20 8.50 7.30 6.65 6.35 7.10 11.60 ilnfo: Information Comp: Comprehension Ari: Arithmetic Simi: Similar ities Voc: Vocabulary Disp: Digit Span Picc: Picture Completion j Blkd: Block Design Obja: Object Assembly Codg: Coding ! 105 In comparisons of sex groups on the WISC subtests males surpassed females on Information, Comprehension, Vocabulary, Picture Completion, Picture Arrangement, Block Design, and Object Assembly, while the fe- i males surpassed the males on the Coding. There was little or no differ ence on Arithmetic, Similarities, and Digit Span. I j Viewing the WISC subtests by race-sex groups, the strongest group on Information was Anglo male (6.90) while the poorest was Mexican fe male (5.35). Anglo males also were highest on Comprehension (8.45) while Anglo females (6.15) were lowest. Anglo males did best on Arith metic (6.90) and Mexican females (5.25) did poorest. Anglo males were highest on Similarities (7.95) and Anglo females were lowest (7.15). Anglo males were highest (7.90) and Mexican females lowest (5.10) on ‘ Vocabulary. Negro females were highest (8.50) and Mexican males lowest (6.75) on Digit Span; Anglo males were highest (9.85) on Picture Com- 'pletion while Negro females were lowest (7.30). Anglo males were high est on Picture Arrangement (8.30) and Negro females were lowest (6.65). j Anglo males were highest on Block Design (8.85) and Negro females were ; lowest (6.35) . On Obj ect Assembly Anglo males (9.95) were highest and j I Negro females (7.10) were lowest. On the Coding Negro females were i ■highest (11.60) while Negro males were the lowest (8.40). In summary, i I the Anglo males were the highest performers in all of the WISC subtests j except Digit Span and Coding, and they were lowest in none of the sub- j tests. Anglo females were lowest in Comprehension and Similarities and I highest in none. Mexican males were neither highest nor lowest on any of the subtests. Mexican females were lowest on Information, Arithmetic, Vocabulary, and Digit Span, and highest on none. Negro males were low est on Coding and highest on none. Negro females were highest on Digit Span and Coding, and lowest on Picture Completion, Picture Arrangement, Block Design and Object Assembly. TABLE 13 WAIS MEAN SUBTEST SCALE SCORES* . INFO COMP ARI SIMI DISP VOCA DISY PICC BLKD PICA OBJA All Subjects 6.93 8.32 6.93 8.50 8.15 7.69 9.33 8.50 8.32 7.70 9.43 Anglo 7.60 8.87 7.65 9.20 8.10 8.30 9.72 9.02 9.10 7.92 9.62 Mexican 6.38 7.68 6.22 7.85 7.60 6.82 8.95 8.45 8.60 7.42 9.68 Negro 6.82 8.42 6.92 8.72 8.75 7.95 9.32 8.02 7.28 7.78 9.00 Male 7.08 8.62 7.35 8.65 8.08 7.85 8.75 8.98 8.85 7.63 9.73 Female 6.78 8.03 6.48 8.53 8.22 7.53 9.92 8.02 7.80 7.80 9.13 Anglo Male 7.80 9.10 8.25 9.60 7.90 9.00 9.05 9.65 9.70 8.05 10.65 Anglo Female 7.40 8.65 7.05 8.80 8.30 7.6010.40 8.40 8.50 7.80 8.60 Mexican Male 7.15 8.00 7.20 7.75 7.85 5.75 8.70 8.80 9.30 7.55 9.50 Mex. Female 5.60 7.35 5.25 7.95 7.35 6.90 9.20 8.10 7.90 7.30 9.85 Negro Male 6.30 8.75 6.60 8.60. 8.50 7.80 8.50 8.50 7.55 7.30 9.05 Negro Female 7.35 8.10 7.25 8.85 9.00 8.1010.15 7.55 7.00 8.25 8.95 i* Converted to Scale scores equivalent to WISC Scale scores with ! Mean=10 and S.D.=3 (Wechsler, 1955, 99-101) i It will be recalled that the WAIS subtest scale scores used in I | Table 13 are transformed to have means and standard deviations equal to those of the WISC (ten and three respectively). According to Table 13 j Ithe subjects in the sample as a whole were almost average on Object As- j sembly and Digit Symbol, and in the normal range on Similarities, 107 Picture Completion, Block Design, Comprehension, and Digit Span. Their . poorest performance was on Information and Arithmetic. In teims of racial groups the Anglos were nearly average on Simi larities, Digit Symbol, Picture Completion, Block Design, and Object Assembly. Their poorest WAIS subtests were Information, Arithmetic, and Picture Completion. The Mexicans were almost average on Object Assembly and in the normal range on Digit Symbol, Picture Completion, and Block Design. Their poorest subtests were Information, Vocabulary, and Arith metic. The Negroes were almost average in Digit Symbol and Object Assembly and in the normal range in Comprehension, Similarities, Digit I jSpan, and Picture Completion. Their poorest subtests were Information j land Arithmetic. i i In terms of race-sex subgroups, Anglo males surpassed the others .in eight of the eleven subtests, but they ranked fourth in Digit Span ; I and Digit Symbol and second in Picture Arrangement. Anglo females ! jranked first in Digit Symbol but last in Object Assembly. Mexican males] Jranked last in Similarities and Vocabulary. Mexican females ranked last ! I ! I I in Information, Comprehension, Arithmetic, and Digit Span and shared | | ' i |last place with Negro males on Picture Arrangement. Negro males were j i ‘ - I j j Ilast on Digit Symbol. Negro females were first on Digit Span and Pic- i ture Arrangement but last on Picture Completion and Block Design. i ! We are interested in the difference between comparable WISC and j . iWAIS subtests. Table 14 presents the differences obtained by subtract- i ing WISC scale score means from the adjusted WAIS subtest scale score I means. The subtests are ranked according to the amount (and per cent) ; 108 of difference contributed to the I.Q.'s. It will be noted that where the WISC subtest value is larger than that of the MIS, a minus value is obtained. TABLE 14 SUBTEST CONTRIBUTIONS TO WMS-MINUS-WISC I.Q. DIFFERENCES FOR ALL SUBJECTS lank Subtest Scale Points Contributed Percent of Total Contribution 1 Vocabulary 1.64 2119 2 Comprehension 1.24 16.6 3 Similarities 1.10 14.7 4 Block Design 1.00 13.4 5 Arithmetic .73 9.7 6 Information .66 8.8 7 Digit Span .65 8.7 8 Object Assembly .61 8.1 9 Picture Arrangement .45 6.0 10 Picture Completion .28 3.7 11 Coding (Digit Symbol) (Minus) .87 (Minus) 11.6 Total 7.51 100.0 « ■ . . . . . . . . . . . . . . . . . . . . . . . . . - — - ■ ■ i i 1 1 ! According to Table 14 the Vocabulary subtest difference in sub- j ; i itest contribution accounted for 21.9 per cent of the differences in Full) | i Scale I.Q. between the two tests. Other major contributors to the Full j Scale differences were Comprehension (16.6)per cent), Similarities : i : j (14.7 per cent), and Block Design (13.4 per cent). All of the MIS sub- ! tests appear to be easier than the comparable WISC subtests for this 109 sample except for Coding, which showed a surprising reversal. It should be noted that for all subtests other than Coding the mean scale scores of this below-average intelligence population were below the mean of the normal population, but the WISC Coding mean was above the norm popula tion mean. Table 15 presents subtest difference contributions by race inclu ding the rank and scale points contributed by each subtest. TABLE 15 SUBTEST CONTRIBUTIONS TO WAIS-MINUS-WISC I.Q. DIFFERENCES BY RACE Subtest ANGLO Scale Rank Points MEXICAN Scale Rank Points NEGRO Scale Rank Points Information 5 .90 7 .38 8 .70 Comprehension 2 1.57 2 1.10 3 1.04 Arithmetic 6 .85 6 .42 4 .90 Similarities 1 1.65 9 .23 2 1.42 Vocabulary 3 1.33 1 1.60 1 2.00 I Digit Span 7 .65 10 .15 9 .65 : Picture Completion 9 .35 5 .48 10 0.00 | Picture Arrangement 10 .30 8 .32 7 .76 I Block Design 4 1.22 3 1.00 6 .78 t . ’ Object Assembly 8 .40 4 .56 5 .88 i ! Coding (Digit Symbol) 11 (-).90 11 (-).68 11 CO.69 I Total 8.32 5.56 8.45 110 For the Anglos the greatest WAIS-minus-WISC subtest disparity was obtained on Similarities, followed by Comprehension, Vocabulary, and Block Design. All subtests showed this WAIS-over-WISC trend except Co ding which was a sharp reversal. For the Mexicans the greatest!',WAIS- minus-WISC subtest disparity was on Vocabulary, followed by Comprehen sion and Block Design. As in the case of the Anglos, the Mexicans reversed the WAIS-over-WISC trend on the Coding. For the Negroes there was a full two .scale point difference on the Vocabulary subtest in fa vor of the WAIS. This was the largest disparity obtained for any sex group on any subtest, and it was followed by Similarities and Compre hension. The Negroes had no difference between WISC and WAIS on the Picture Completion, but their Coding reversal was the same size as that jof the Mexicans. Comparing the races on subtest differences, we find |that while the Anglo and Negro groups had large WAIS-minus-WISC dispar ities, on the Similarities there was relatively little difference for : I ; the Mexicans. While the Vocabulary disparity was very large for all j three races, it was much greater for the Negroes than for the Anglos. j j There was very little difference between the WISC and WAIS Digit Span j ; subtest performance by the Mexicans, but the Anglos and Negroes had j : sizable differences in favor of the WAIS. The sample from population number two consisted of 380 case stu- ; idies of suspected mental retardates who, subsequent to the testing, had i 'either graduated or dropped out of school. Of the 380 cases in popu- | i lation two, 304 had been administered the WISC and the remaining 76 had j been administered the WAIS. Table 16 presents a comparison of the in- i Ill telligence quotients and reading achievement scores for subjects taking the WISC and those taking the WAIS. TABLE 16 READING ACHIEVEMENT AND I.Q. 's OF SAMPLE TWO WISC Mean S.D. WAIS Mean S.D. Age 13.63 1.10 16.42 0.63 Verbal I.Q. 73.25 10.09 81.89 9:46 Performance I.Q. 77.15 14.15 82.59 9.73 Full Scale I.Q. 72.91 11.61 80.98 8.87 Reading Level ‘ 3.14 2.39 4.56 2.05 Actual Grade Placement 8.67 1.14 11.01 0.65 Reading Index 0.36 0.26 0.41 0.18 N (304) (76) ■ According to Table 16 the average student in this population of j suspected mental retardates was "on schedule" in terms of chronological i i ;age and actual grade placement at the time he was tested with the WISC, j ibut the average student was about three months behind schedule at the j i i |time of taking the WAIS. Although it cannot be maintained that the | ‘ ' j |sample of suspected retardates who took the WAIS is comparable to the i [ j ! sample of suspected retardates who took the WISC, it is noteworthy that j i j | there are differences between the WISC and WAIS Verbal, Performance and | |Full Scales of 8.64, 5,44 and 8.07 respectively. In each case the WAIS ' 112 surpasses the WISC and Performance surpasses Verbal. At the time of testing the younger group which took the WISC had a mean reading index of .36; on the other hand the older group which took the WAIS had a slightly higher reading index (.41). In the sample from population number one (which was composed of students obtaining I.Q.'s below 90) every subject took both the WISC and the WAIS. In the sample from population number two (which was composed of suspected mental retardates) every subject took either the WISC or the WAIS, but no subject took both tests. As expected, mean I.Q.'s of sample one are higher than those for sample two, yet there is a simi larity between the WAIS-minus-WISC differences of the two samples. Table 17 presents the comparative data for the two samples. | TABLE 17 COMPARISON OF THE I.Q. 's OF SAMPLE ONE AND SAMPLE TWO ! SAMPLE 1 SAMPLE 2 i WISC I (Verbal I.Q. i (S.D.) iPerformance I.Q. | (S.D.) Full Scale I.Q. (S.D.) WISC WAIS DIFF WISC WAIS DIFF 80.11 87.83 7.72 73.25 81.79 8.64 10.76 9.49 10.09 9.46 88.51 90.95 2.44 77.15 82.59 5.44 13.96 9.98 14.15 9.73 82.52 88.80 6.28 72; 91 80.98 8.07 11.17 8.61 11.61 8.87 In sample one and sample two the WAIS-minus-WISC differences are i t considerably greater on the Verbal scales than on the Performance scales. The WISC and WAIS Verbal scale standard deviations are comparable in the! | two samples, but the WISC Performance scale standard deviations are much 113 larger than those of the WAIS Performance. Table 18 presents a comparison of the intelligence and reading test scores of the regular class students and the EMR class students from the sample frcm population number two. TABLE 18 A COMPARISON OF THE REGULAR AND EMR CLASS STUDENTS IN SAMPLE TWO REGULAR Mean S.D. EMR Mean S.D. Age at Testing 14.12 1.62 14.29 1.38 Verbal I.Q. 80.56 9.09 68.24 7.92 Performance I.Q. 84.90 11.06 70.19 11.82 1 Full Scale I.Q. 80.93 9.23 66.79 9.14 Reading Level 3.89 1.86 2.70 1.43 Actual Grade Placement 9.11 1.54 9.18 1.26 Reading Index 0.42 0.19 0.29 0.14 N i (208) (172) The mean Verbal and Full Scale I.Q.'s of the subjects in sample j number two was a little below 81. With a standard deviation of more i than nine, this indicates that there were a number of students who qual- i | ified for EMR placement on the basis of I.Q. but who, for one reason or : another, remained in regular classes. There was a fourteen point dif- j ference between the means of the EMR and regular class Full Scale I.Q.'s, The reading indexes of the regular class and EMR class groups were .42 114 and .29 respectively. Table 19 presents a comparison of graduates and dropouts of stu dents in sample number two of regular class placement. TABLE 19 A COMPARISON OF GRADUATES AND DROPOUTS IN REGULAR CLASS PLACEMENT GRADUATES DROPOUTS Mean S.D. Mean.' S.D. Age at Testing • 14.07 1.74 14.15 1.54 Verbal I.Q. 83.55 9.20 78.49 8.44 Performance I.Q. 85.24 12.15 84.67 10.28 Full Scale I.Q. 82.87 9.66 79.58 8.71 Actual Grade Placement 9.19 1.61 9.06 1.50 Reading Level 4.89 1.99 3.20 1.41 Reading Index 0.53 0.19 0.35 0.14 According to Table 19 there were differences of 5.00, 0.57, and 13.29 between graduates and dropouts on the Verbal, Performance, and | iFull Scales respectively. With standard deviations ranging between I |eight and twelve, this suggests a great deal of overlap between the success and failure distributions. The Performance scale appears to have no discriminating ability whatever in this sample. There is a I somewhat better discrimination on the part of the reading index. ; Hypothesis number two suggests that individual reading achieve- ment test scores will predict more accurately than individual intelli- 115 gence tests with reference to success in regular classes for suspected mental retardates. To test this hypothesis biserial correlations were computed between each of four continuous variables (Verbal I.Q., Per formance I.Q., Full Scale I.Q., and Reading Index) and the discrete variable: graduate versus non-graduate. Table 20 presents these cor relation values along with the levels of significance for both regular and EMR class students. TABLE 20 I.Q. AND READING INDEX CORRELATIONS WITH GRADUATE STATUS Type Class r Number Required for Significance Verbal I.Q. Performance I.Q. Full Scale Reading Index Regular 208 .05=.170 .01=.224 .349** .031 .223* .602** EMR 172 .05=.189 .01=.246 .194* .188 .229* .328** = Significant at .05 but not .01 level I j * * = Significant at .01 level j 1 ■ i I } i ; According to Table 20 the correlation (.349) between Verbal I.Q. I i land graduation in regular classes for suspected mental retardates is j ■significant at the .01 level. The correlation (.194) between Verbal i |I.Q. and remaining in EMR classes until graduation is significant at ithe .05 level. Performance I.Q. correlations with graduation do not | ; reach significance for either regular or EMR classes, although the | I ' I •value for the EMR group (.188) missed significance by only .001. Cor- j relations between Full Scale I.Q. and graduation were significant at 116 the .05 level in both regular and EMR classes. Correlations between the reading index and graduation were significant far beyond the . 01 level for both regular and EMR class students. In order to determine whether the apparent superiority of the reading index over the I.Q. scales was significant, the correlation values were converted into Fisher's z function and the significance of the difference between z values was determined using the formula: 5 D * = ^ + (Standard Error of the difference z coefficients: Garrett, 1961, 241.) The computational procedure for determining the level of signifi cance of the differences between I.Q. and the Reading Index correlations with graduate status for regular class students is presented as Appendix |D, Exhibits 1 and 2. Table 21 and Table 22 summarize the results of ; i 'these computations. | i TABLE 21 | SIGNIFICANCE OF THE DIFFERENCES BETWEEN I.Q. AND READING INDEX CORRELATIONS WITH GRADUATE STATUS FOR REGULAR CLASS STUDENTS 1 Verbal I.Q. Performance I.Q. Full Scale I.Q. Reading Index 1 Correlation Value .349 .031 .223 .602 iFisher Value of z .37 .03 .22 .69 |Result of Comparison Index VIQ Index PIQ Index FSIQ i i p<\001 p<.001 p<.001 117 According to Table 21 the superiority of the reading index over each of the three I.Q. scales as a predictor of success in regular classes is significant beyond the .001 level. TABLE 22 SIGNIFICANCE OF THE DIFFERENCES BETWEEN I.Q. AND READING INDEX CORRELATIONS WITH GRADUATE STATUS FOR EMR STUDENTS Verbal I.Q. Performance I.Q. Full Scale I.Q. Reading Index Correlation Value Fisher z Result of Comparison .194 .19 Index VIQ (Not Signif) .188 .19 Index PIQ (Not Signif) .229 .23 Index FSIQ (Not Signif) .328 .34 Although the correlation between the reading index and graduation I in EMR classes is greater than the correlations between the three I.Q. ; scales and graduation in EMR classes, the differences do not reach a i significant (.05) level. i | In order to make more detailed comparisons, the I.Q.’s and read ing index values were arbitrarily dichotomized into distributions of | i I "correct" and "incorrect" predictions. Table 23 summarizes these data j I for the 208.regular class subjects from the sample number two. j I j Table 23 indicates that the most accurate predictions of success j 1 • ! j in graduating in regular classes was achieved by the reading index j * ' ! !(74 per cent) followed by the WISC Verbal Scale (65 per cent). The ; WAIS Verbal and Full Scale predicted with 56 per cent accuracy, the j 'WAIS Performance with 52 per cent, and the WISC Performance with 44 118 per cent accuracy. It should be kept in mind that chance accuracy in this case would be 50 per cent. TABLE 23 I.Q. VERSUS READING INDEX AS PREDICTORS OF SUCCESS IN REGULAR CLASSES Verb. I.Q. Perf. I.Q. FS I.Q. Verb. I.Q. Perf. I.Q. FS I.Q. Reading Index Number of Subjects 158 158 158 50 50 50 208 Correct Predictions 103 70 79 28 26 28 154 Percent Correct 65.2 44.3 50.0 56.0 52.0 56.0 74.0 According to Table 24, when the results from administration of the WISC and the WAIS are combined the accuracy of prediction of the Full Scales is only 1.4 per cent better than chance. On the other hand | the reading index predictions are 24 per cent better than chance. j i TABLE 24 j ! COMBINED WISC AND WAIS FULL SCALE | VERSUS READING INDEX i I.Q. Tests Reading Index Number of Subjects 208 208 Correct Predictions 107 154 Per cent Correct 51.4 74.0 j The significance of the difference in accuracy of prediction be- j j J tween the Full Scale I.Q.'s of the WISC and WAIS combined and the read- I i i . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 ing index was computed using the procedure described in Chapter III and the results are summarized here: (1) Standard error of a percentage Formula: N Full Scale I.Q. Reading Index (51.4) (48.6) 1 | (74) (26) 208 | 208 = 3.46 = 3.01 (2) Standard error of a difference between percentages *P1 + ^ 2 = 4,58 (3) p, “ Po 1 2 = 5.15 (Significant .001) On the basis of the above data the superiority of the reading index over the Full Scale I.Q. is highly significant, j Table 25 presents a breakdown of the I.Q.'s into ranges of ten points each along with the number and per cent of students graduating I ' | In regular classes. Table 26 presents the same data for the EMR class students. According to Table 25, students who are tested as suspected mental j retardates and obtain Full Scale I.Q.'s of 90 or above have a two-out-of- 120 three chance of graduating in regular classes. However, for those who score below 90, the chances of graduating in regular classes appear to be only a little better than one-out-of-three, and the odds do not seem to be better for those in the upper part of this range than for those in the lower part. TABLE 25 STUDENTS GRADUATING IN REGULAR CLASSES BY I.Q. RANGES I.Q. Range Number of Students Number Graduating Percent 90 and above 34 23 67.7 80-89 96 34 35.4 70-79 56 20 35.7 60-69 17 6 35.3 50-59 5 2 40.0 TABLE 26 STUDENTS GRADUATING IN EMR CLASSES BY I.Q. RANGES I.Q. RANGE Number of Students Number Graduating Percent 80-89 10 5 50.0 70-79 64 44 68.7 60-69 61 32 52.4 50-59 44 14 31.8 Below 50 6 3 50,0 Of the ten students who were placed in EMR classes despite Full I i | Scale I.Q.'s in the dull normal I.Q. range only half managed to gradu- | ‘ate. Students in the 70 to 79 I.Q. range (sometimes referred to as ! borderline EMR) had better than a two-out-of-three chance of graduating 121 in EMR classes. Subjects with I.Q.'s in the sixties and fifties had decreasing chances of graduating. Subjects with I.Q.'s low enough to qualify for special classes for trainable mentally retarded (IMR) had an even chance of graduating in EMR classes, but the number of subjects in that category was quite small. It may be observed in comparing Table 25 to Table 26 that increasing I.Q. within the 50 to 89 range is jrelated to increased chance of graduating in EMR classes but not in regular classes. Figure 1 presents a comparison by I.Q. ranges of students gradu ating in regular class placement and students graduating in EMR placement. COMPARISON OF I.Q. 's OF REGULAR AND EMR CLASS GRADUATES ; PERCENT GRADUATING FIGURE 1 80 70 60 50 40 30 20 10 0- 50 FULL SCALE 49 59 I.Q. = EMR Class Graduates 122 In an effort to study the relationship between reading index and success in graduating from high school, Tables 26 and 27 were construc ted showing the percent of graduates by reading index for regular and EMR class students respectively. TABLE 27 STUDENTS GRADUATING IN REGULAR CLASSES BY READING INDEX RANGES Index Range Number of Students Number of Graduates Percent .900 and above 5 4 80.0 .800-.899 4 3 75.0 .700-.799 11 10 90.9 .600-.699 18 14 77.8 .500-.599 23 15 65.2 .400-.499 43 16 37.2 .300-.399 51 15 29.4 .200-.299 33 7 21.2 .100-.199 18 1 5.6 .000-.099 2 0 0.0 Total 208 85 40.9 .500 and above 61 46 75.4 .000-.499 147 39 26.5 It will be recalled that the reading index expresses the rela tionship between reading achievement level and actual grade placement 123 and that a student with average reading ability would have an index of 1.00. Table 27 shows a consistent increase in chance of success in reg ular classes with increase in reading index until a high of 90.9 per cent is achieved for those with reading indices between .700 and .799. It would seem that students who have been referred for testing as sus pected mental retardates, and whose reading ability were as good as 80 per cent of grade level, have probably been referred for reasons I other than intelligence or basic reading ability. TABLE 28 STUDENTS GRADUATING IN EMR CLASSES BY READING INDEX RANGES i Index Range Number of Students Number of Graduates Percent 1.800 and above j 0 0 _ ~ .700-.799 2 2 100.0 .600-.699 3 2 66.7 .500-.599 j 9 6 66.7 j.400-.499 [ 17 10 58.8 j.300-.399 42 33 78.6 .200-.299 52 25 48.1 .100-.199 i 41 19 46.3 .000-.099 6 1 16.7 1 Total 172 98 56.9 .500 and above 14 10 71.4 .000-.499 158 88 55.7 124 According to Table 28, reading ability appears to be related to success in graduating from EMR classes although the relationship is not as marked as it is in regular class students. In the cases where sub jects have reading indices above .300 the odds are better than seven-out- of-ten for successful graduation in special training classes. Between .100 and .299 the odds are slightly less than even and for those below .100 the chances are only one in six. In order to show the relationship between Reading Index and suc cess in graduating for both regular class and EMR class students, Figure 2 presents the data from Tables 27 and 28 in graphic form. FIGURE 2 COMPARISON OF READING INDICES OF REGULAR AND EMR CLASS GRADUATES I PERCENT GRADUATING 100 READING ; INDEX .000- .100- .200- .300- .400- .500- .600- .700- .800- .900 .099 .199 .299 .399 .499 .599 .699 .799 .899 § upj = Regular Class Students ....... EMR Class Students CHAPTER V DISCUSSION, CONCLUSIONS, AND RECOMMENDATIONS Discussion For the purpose of the present investigation two hypotheses were itested: 1. WAIS I.Q.'s will exceed WISC I.Q.'s for the population of be low-average intelligence subjects as a whole, for all race, sex, and race-sex groups on the Verbal Scale, Performance Scale, and Full Scale. 2. Individual reading achievement tests will predict more accu rately than individual I.Q. tests with reference to success in gradua tion from high school in regular classes by suspected mental retardates. ; the tests of hypothesis number one were based on data obtained i 'from a sample of one hundred twenty sixteen year-old students who were below average intelligence. The sample contained twenty males and 'twenty females from each of the Anglo-American, Mexican-American, and iNegro-American race groups. ! The tests of hypothesis number two were based on data obtained from a sample of 380 junior and senior high school students who had ibeen referred to the district psychological services, had been tested ; i |as suspected mental retardates, and for whom it could be established j Ithat they had either graduated or had dropped out of school without |transferring to another. Of the 380 students making up sample number j two, 208 were in regular classes, and of these 208, 85 graduated and ! 125 126 74 dropped out. Analysis of variance was used to test hypothesis number one. Bi serial correlation coefficients related to the second hypothesis were computed and these values were converted into Fisher's z function and the significance of the difference between z values was determined by formula as described in Chapter III. i 1 | Test of Hypothesis One i j The purpose of hypothesis number one was to determine if the WAIS would produce higher I.Q.'s than the WISC if the two tests were admin istered to the same subjects under controlled conditions. Present de cision-making by psychologists is based on an assumption that the two j itests are comparable. Information as to the size and significance of ! i ' i differences between the two tests and knowledge of the influence or j race and sex on these differences could contribute to more effective | use of the WISC andtlie:WAIS in important decision-making with reference j ito educational planning and placement of suspected mental retardates. J ! i Entire Sample j I The analysis of variance indicates that the overall WAIS-minus- j IWISC I.Q. differences are significant beyond the .001 level for the Ver-j 1 ' j -bal and Full Scale and beyond the .01 level for the Performance Scale, j jRace Groups j | ! ! Among the race groups there is the widest disparity between the ; I | Itests for the Negroes, who average more than eight and one-half I.Q. j ; | I ; ! points higher on the WAIS than the WISC Verbal Scale and nearly eight ; jI.Q. points higher on the Full Scale--which is the most influential of 127 the three scales for placement purposes. This means, in effect, that a Negro who is placed in EMR classes on the basis of a WISC Full Scale I.Q. between 72 and 79 would have scored above the EMR range on the WAIS. Anglo subjects have a six point disparity and Mexicans have a five point disparity on the Full Scale, both in favor of the WAIS. The differences between the sexes were in favor of a larger WAIS-minus-WISC disparity i f or females (8.41 I.Q. points) than for the males (6.98), but this dif ference did not reach the .05 level of significance. Correlations | For the sample as a whole, correlations (see Table 10) between the i 'comparable scales of the two tests are moderately high but not as high as those obtained by Webb (1963 and 1964) with much smaller samples of | public school special class retardates. In his study of Negro EMR's in 1963 he reported correlations of *80, .91, and .84; and a study of Anglo and Negro EMR's in 1964 showed correlations of .87, .83, and .91 i for Verbal, Performance and Full Scales respectively. j ! I ! For the race groups, the Negroes' Verbal Scales of the two tests correlated very high (.89), yet the Negroes had the largest WAIS-minus- ! ; j WISC mean difference on the Verbal Scale (8.95). In general, the Mexi- ; j ' ! Jean subjects tended to have the lowest correlations between the two j i ! .tests. This appears to be related to the fact that their scores tended 'to be less scattered than the other races. The standard deviations for | I jthe Mexicans were noticeably smaller than for the other two races. j j | Effect on Placement | According to Table 6, 25.8, 37.5, and 16.7 per cent of the sub- 128 jects in this sample would be placed in EMR classes on the basis of the WISC Full Scale, Verbal, and Performance I.Q.'s respectively, but placed in regular classes on the basis of WAIS I.Q.'s. Since in practice the Full Scale is usually the one given the most consideration, it would appear that one-fourth of the students who would qualify for EMR place ment on the basis of the WISC do not qualify on the WAIS. The reverse of this did not hold in a single case where the Full Scale was concerned but there was one case on the Verbal and two on the Performance where the trend was reversed. Subtest Contribution Table 12 was constructed showing the mean subtest scale scores for the WISC and Table 13 was constructed showing these values for the WAIS | iin an effort to determine which subtests, if any, accounted for the | I large differences between the two tests. The most noteworthy phenomenon : ! to be observed in the WISC subtest scores is the large contribution to j I ;the Full Scale I.Q.'s of the Coding subtest as compared to all of the \ i I j others. This was true for all three race groups. Below-average intel- j I I jligence subjects of all three race groups appear to be handicapped most ! I ' i ; in Vocabulary, Arithmetic, and Information. In addition, the Mexican j I subjects scored low in Comprehension and the Negroes scored low in ! Block Design. : | In the case of the WAIS, the Digit Symbol (which is the equivalent of the WISC Coding) accounted for the greatest contribution to the Full 'Scale I.Q. for the Anglos and the Negroes, but the degree of contribu- 1 tion was less than in the case of the WISC Coding. For the Mexicans, 129 the Digit Symbol was high, but it was surpassed a little by Object Assembly. The Anglos were relatively strong (scale score 9.0 or higher) on Digit Symbol, Object Assembly, Similarities, Block Design, and Picture Completion, and relatively weak (scale score below 8.0) on Information, Arithmetic, and Picture Arrangement. The Mexicans were relatively j strong on Object Assembly and relatively weak on Arithmetic, Informa tion, Vocabulary, Picture Arrangement, Digit Span, Comprehension, and Similarities. The Negroes were relatively strong on Digit Symbol and Object Assembly and relatively weak on Information, Arithmetic, Block Design, Picture Arrangement, and Vocabulary. The next question to be asked related to the differences in sub- test contribution to Full Scale I.Q. between the comparable subtests of j ‘ I : the two scales. Incother words, we wanted to know whether the same sub tests contributed heavily to the Full Scale WISC I.Q.'s as to the Full j Scale WAIS I.Q.'s. According to Table 14 the Vocabulary subtest differ• • i * ! j ; ences in contribution accounted for 21.9 per cent of the differences in; jl.Q. between the two tests. Other major contributors to the WISC-WAIS ; | ; differences were Comprehension (16.6 per cent), Similarities (14.7), j land Block Design (13.4). All of the WAIS subtests were easier than the: I |WISC comparable subtests for this sample except for the Coding, which showed a surprising reversal which the investigator is unable to ex- ; j i 'plain. It should be noted that for allcother subtests the mean scale j j I ! scores of this below-average intelligence population were below the j average for the norm population, but the WISC Coding was above the norm 130 population mean. For Anglo subjects the largest contributor to differences in Full Scale I.Q.'s on the two tests was Similarities, closely followed by Comprehension and Vocabulary. The Anglo subjects did not depart very far from the general trends, although they had the largest WlSC-over- WAIS reversal in the Coding subtest. The Mexican subjects had their greatest WAIS-WISC subtest differ ence on Vocabulary. Whereas the Anglos and Negroes had large subtest differences between WISC and WAIS on Similarities, the Mexican differ ence was very small. The most noteworthy aspect of the Negro between-subtest compari sons was the 2.00 scale point difference between the WISC and WAIS I | Vocabulary subtests which was the largest difference for any race group: ! ' ’ I Negro subj ects had identical subtest means on Picture Completion of WISC and WAIS, and the WAIS-over-WISC advantage on Object Assembly and j Picture Arrangement was somewhat greater in the case of the Anglos and | Mexicans. i i | Test of Hypothesis Two i ! I The purpose of hypothesis number two was to determine whether in-j j i I dividual intelligence tests are useful in predicting whether a suspected ! i |mental retardate will survive regular classes on the basis of his I.Q. i Three methods were used in evaluating the I.Q. tests in this regard, j (1) Hypothesis number two was tested by computing Biserial Correlation i Coefficients between I.Q. and graduation and between reading achievement i and graduation. Then the significance of the difference between 131 correlation coefficients was obtained. (2) The per cent of accuracy with which I.Q. predicted success in regular classes for the present sample was computed and this value was compared to that of chance (50 per cent). Here the question was the amount of improvement over chance provided by I.Q. predictions. (3) The predictive accuracy of the I.Q. tests was compared to that of the individual reading tests. ! It may be recalled that the sample from population number one was composed of students scoring below the average range on an individual or group I.Q. test but who were not necessarily doing badly academically. On the other hand, the sample from population number two was composed entirely of students who were failing academically and who were in ad- j dition suspected of being mentally retarded. Therefore it is to be | i ■expected that the mean I.Q.'s for sample two would be lower than those |for sample one, and this proved to be the case. According to the data 'in Table 17, in both samples one and two the WAIS Verbal, Performance, j land Full Scale I.Q.'s surpass the comparable WISC scales. Moreover, in | i both sample one and sample two the WAIS-minus-WISC differences are con- ! ! j isiderably greater on the Verbal Scales than on the Performance Scales. ! I j ;The WISC and WAIS Verbal standard deviations are comparable in the two I i j i samples, but the WISC Performance Scale standard deviations are much i ; ! I I larger than those of the WAIS Performance. Comparing the data from ! I sample one to that of sample two (Table 17) suggests that the further below the population mean of 100, the greater the disparity between the comparable WISC and WAIS scales. This holds true for all three scales. ; . i Hyopthesis number two was tested by the use of Biserial Correia- 132 tions between each of the four predictors (Verbal I.Q., Performance I.Q., Full Scale I.Q., and reading index) and the dependent variable, gradua tion status (graduate or dropout). The data presented in Table 20 indicate that the correlations with graduate status are .349 for Verbal I.Q., .031 for Performance I.Q., .223 for Full Scale I.Q.-, and .602 for the Reading Index. The preceding values apply to students in regular classes. Although the Verbal correlation was found to be significant beyond the .05 level, it is too low to be of much practical value in determining class placement for suspected individual retardates. According to Table 20 the Perfoimance I.Q. is of no value whatever in predicting success in regular classes. Using the Verbal I.Q. alone would be much more effective. The correlation between graduation and reading index was far above those of the I.Q.'s. It will be noted that significant,, but low, positive correlations were obtained between gra duation for EMR students and Verbal I.Q. (.194, significant beyond .05), Full Scale I.Q. (.229, significant beyond .05), and reading index (.328, significant beyond the .01 level, but none of these correlations was high enough to be useful in iindividual placement decisions. Although still not significant, the correlation between graduation and Perform ance I.Q. was higher for EMR students than it was for regular class stu dents . The correlations for the regular class students were tested for the significance of the differences between the I.Q.'s and the reading index by converting the correlation values to the Fisher's z function and applying the formula for obtaining the standard error of the dif ference between z coefficients as shown in Chapter 4. The difference 133 was found to be significant beyond the .001 level in favor of the read- index. In terms of the comparison of I.Q. and reading test as predic tors of academic success for suspected mental retardates, the reading test is far superior to the intelligence test and the research hypothe sis is supported. Although it was not a part of the second hypothesis, the correla tions between graduate status and each of the three predictors were compared for possible significant differences (Table 22) for the EMR students as was done in the case of the regular class students. In each case the reading index was found to be somewhat superior to the I.Q. tests but none of the comparisons produced significant differences (.05). j • The preceding analysis of the correlations between I.Q. and read ying index and graduate status has shown that successful students tend ; to have higher intelligence and reading ability than do unsuccessful j ; students, but it left unclear the extent to which the predictors could I I be depended upon to discriminate between students who are likely to j |succeed and those likely to fail in regular classes. Table 23 indi- | : ] i I cates the number and per cent of correct predictions made by the I.Q.'s | : ! i i j and reading indices using the previously described (I.Q. 80 and reading j ! | ; index .50) cutoff points and comparing the predictions to the outcomes i in terms of graduation or dropout. Table 23 shows predictive accuracy ! percentages as follows: WISC Verbal 65.2; Perfoimance 44.3; and Full I ! ' i ;Scale 50.0. WAIS Verbal 56.0; Performance 52.0; and Full Scale 56.0. , i Reading index 74.0. Using "chance" as a reference point, the WISC 134 Perfoimance is 6 per cent worse than chance; and WISC Full Scale I.Q. is exactly the same as chance. The three WAIS scales are six, two, and six per cent better than chance for the Verbal, Performance, and Full Scale I.Q.'s respectively. On the other hand, the reading index pre dicted with 74 per cent accuracy, 24 per cent better than chance. The combined WISC and WAIS predictions were compared to those of the read ing index for the 208 subjects in sample two and the significance of the difference between the accuracy of the Full Scale I.Q. (51.4 per cent) and the reading index (74.0 per cent) was computed as described in Chapter 4. The difference between the predicting accuracy of the read ing index and that of the Full Scale I.Q. was significant beyond the .001 level. On the basis of this comparison the superiority of the reading index over the Full Scale I.Q. is highly significant. Table 25 shows the per cent of students in regular classes who graduated by ranges of ten I.Q. points each. We would expect that at ithe low I.Q. ranges the per cent of students able to graduate in regu- j ! lar classes would be almost zero and that this per cent would increase j ; i ; rather gradually as the normal range was approached. However, accord- j• ! ing to Table 25, for suspected mental retardates, there is hardly any I i ' I | relationship between success in regular classes and I.Q. Two out of ! i five students with I.Q.'s between 50 and 59 were able to graduate in i i ' j ! regular classes, but of course five is a very small number on which to I ! i | base a generalization. It is interesting to note that there was no i difference in the percentage of students who managed to graduate in I ; regular classes with I.Q.'s between 60 and 69 as opposed to the per- j . . 135 centage who graduated with I.Q.'s in the next two higher ranges: 70 to 79 and 80 to 89. It was 35 per cent in each case. It may be that successful graduation in regular classes is not based on intelligence for students below the normal range. Above the noimal range the per cent graduating was 67.7. There appears to be a closer relationship between I.Q. and grad uating in EMR classes than there is in regular classes. In our sample were ten students with I.Q.'s in the dull noimal range yet who were placed in EMR classes. Half of them managed to graduate in EMR classes. At the other end of the scale were six students with I.Q.'s below 50 (and therefore in terms of I.Q. probably eligible for special classes for the trainable mentally retarded), and half of them graduated. If we eliminate the dull noimals and the possible TMR's from our EMR sample, !the remaining students who fall in the EMR range (50-79) do show a fairly strong relationship between I.Q. and probability of success in graduating in EMR classes. Those in the 50-59 I.Q. range had a 31.8 ' i per cent graduation rate. Those in the 60-69 I.Q. range had a 52.4 per i ! jcent graduation rate; and those in the upper EMR range (70-79) had a ;68.7 per cent graduation rate. (See Figure 1) i I j The relationship between reading ability and graduation in regular i ' classes is shown to be high (Table 27) when the ranges of the reading index are given with percent of graduates. Only two students were found in regular classes whose reading achievement was less than one-tenth their actual grade placement, and both of them failed to graduate. Of |18 students whose reading ability was between one-tenth and one-fifth | 136 of their actual grade level one (5.6 per cent) graduated in regular classes. For the data for the 208 subjects in regular classes in sample two, every increase in index range was accompanied by an increase in per cent of those graduating. This was true until 80 per cent of grade level reading ability was reached. After that point the cell size was small and the data inconclusive. It may be that other factors, such as environmental variables, motivation, or personality exert the dominant influences which deteimine graduation or non-graduation above that point. It should be kept in mind that students with near-average reading ability would not ordinarily be referred for testing as suspected mental retardates. Conclusions j Generalizations based upon the results of this study must be made with caution because of the nature of the sample populations. Popula tion number one was composed of students who had scored below the aver- j | age range on a group or individual intelligence test. Level of academic > • ' I i i Iperfoimance and type of class placement (regular or EMR classes) did not influence their selection. Population number two was made up of stu- j i 1 dents who had been referred for testing as suspected retardates. The i I ' i ; findings that were statistically significant as well as some trends j 1 j which were apparent, however, suggested that certain conclusions are j I tenable. These are listed in the order in whidi they appeared in the | (body of the dissertation: | j ! 1. Students who are below the average range in intelligence J obtained significantly higher WAIS than WISC I.Q.'s. The 137 discrepancy was greatest on the Verbal Scale, followed by the Full Scale, and least on the Performance Scale. 2. The disparity between the WISC and WAIS I.Q.’s was signi ficantly greater for the Negroes than for the Anglos and Mexicans. 3. On seventy-five percent of the placement recommendations the WISC and WAIS Full Scales made the same choice, but on twenty-five per cent of the cases the WISC recommended EMR placement while the WAIS recommended regular classes. 4. The Vocabulary, Comprehension, Similarities, and Block Design subtests (in that order) accounted for the largest portion of the differences between the WISC and WAIS I.Q.'s, but all of the sub test pairs made positive contributions except the Coding-Digit Symbol pair which reversed the WAIS- j I ! surpassing-WISC trend. j 5. For all practical purposes the I.Q.'s were found to be ! j | | of little value as predictors of success in high school i ! ' I graduation for suspected retardates, but the reading index ! was found to have considerable predictive value. ! i ; I Recommendations j | Whether the findings reported in this study may be generalized to ! |the larger population of high school students throughout California, or I j throughout the United States, depends upon the representativeness of the j j samples studied. Probably the two samples in this study are not 138 entirely representative of school children of all ages and under the « usual testing conditions. In the sample from the first population the subjects were all within three months of their sixteenth birthdays at which time they took both tests on the same day. This is an artificial situation which would not be duplicated in actual practice. It is not known what the effects were of taking both tests in a single administra tion. It is also not known what were the effects of the knowledge that the testing was being done for the enlightenment of the test adminis- . trator rather than as a result of doing badly in school subjects. Some of the subjects may have secretly doubted the reason for testing given by the examiner. In order to reduce practice effects the subtests of both instru- ments were presented in random order, whereas the norms of the tests i j iwere established with presentation in the conventional sequence. The I ! effect of this variation in test administration cannot be determined. All of the preceding sources of error were accepted reluctantly in i [preference to what appeared to be greater error presented by the alter- |native. ! • ; ' A matter of concern in the present study was the inability to ob- j | i | tain the services of at least one Mexican and one Negro psychologist, j | I Research cited in the review of the literature suggested that race dif- ! ferences between the subject and examiner influence decisions concern- i j |ing class placement as well as the response attitude of the subject. I * I j The present study was confined to subjects with below-average in telligence. It is not known whether the WISC-WAIS differences in: Full 139 Scale I.Q. and in subtest contribution would obtain with average and above-average intelligence subjects. Intelligence was compared to reading achievement as predictors of academic success. No attempt was made to detennine a multiple relation ship between two or more of the variables and the criterion measure which might surpass any one predictor used alone. The correlation values obtained indicated that considerable variance remained unaccoun ted for. In view of these considerations the following recommendations for further research are made: 1. A comparison should be made between the practice effects of counterbalancing the WISC-WAIS order of administration and ran domizing of the twenty-two subtests. j 2. Further study should be made of the effects of race differ ences between subject and examiner. j 3. A comparison should be made between the WISC and WAIS I.Q. I i differences and subtest contributions of average and above- ; average intelligence individuals. i 4. Studies should be undertaken to identify other factors i which may account for some of the variance in predicting ! I academic success. 5. Multiple correlation studies should be undertaken which I combine the predictive values of academic success of the | I Verbal intelligence, reading index, and other variables (as | I for example, Wolf's environmental factors) in an effort to 140 increase the overall predictive accuracy for high school success. The data of the present study indicates that below-average intel ligence subjects will get higher WAIS than WISC I.Q.'s. Reading achievement tests provide fairly good predictions of the chance of suc cess in regular high school classes while I.Q. values do not. On the basis of the findings in this study the following general recommenda tions are made: 1. In view of the discrepancy between the WISC and WAIS the two tests should no longer be considered as comparable for below-average subjects. When comparisons of the two tests are made it is suggested that five, six, and eight points be added to the Full Scale WISC I.Q.'s to obtain the WAIS equivalents j for Mexican, Anglo, and Negro subjects respectively. j 2. The present concept of EMR classes for minority groups should I be replaced by some structure which will remove the stigma j i which is now felt at such placement. ! i 3. Reading ability is so much more closely related to success j in regular classes than I.Q. scores that a reading achievement ! I test score should be given at least equal weight with the I.Q. ! in arriving at decisions involving educational placement. Hopefully this study has cast some additional light on some as pects of the relationship between the WISC and the WAIS. It has con firmed the findings of previous investigators that the WAIS yields \ i higher I.Q.'s than the WISC for below-average intelligence Anglos and i 141 Negroes and has expanded the comparison to include Mexican-Americans. It has also made a small contribution toward the understanding of the relative predictive ability of these two intelligence tests as com pared to individual reading achievement tests. The investigator regards these results as stimuli toward further research in this important area and he hopes that it will be of some benefit to other interested researchers. CHAPTER VI SUMMARY The two major objectives of this study were (1) to investigate the comparability of the Wechsler Intelligence Scale for Children and the Wechsler Adult Intelligence Scale, and (2) to compare the ability of intelligence tests and reading achievement tests to predict gradua tion in regular high school classes for suspected mental retardates. Specifically, the investigator sought answers to the following ques tions: i i 1. Are there significant differences between the WISC and WAIS Ver- I S ' i jbal, Performance, and Full Scale I.Q.'s of below average intelligence j I i Anglo, Mexican-American, and Negro-American boys and. girls? j | I 1 2. What is the practical effect on class placement of any difference j between the two tests? j : I ;3. Which subtests account for the difference in Full Scale I.Q. be- i |tween the two tests? j _ i | 4. How do individual intelligence tests compare to individual read- j ! j i ing achievement tests as predictors of successful graduation in regular j |high school classes for suspected mental retardates? ! ! 1 i ■ i The two samples employed in the study were obtained from the j ! I 1 |Compton Union High School District in Los Angeles, California. The j j j ! literature was explored and pertinent studies were reviewed. Two ■hypotheses were established. The first concerned the comparison of | ! 142 143 the WISC and the WAIS Verbal, Performance, and Pull Scale I.Q.'s for the six race-sex: groups (Anglo male, Anglo Female, Mexican male, Mexi can female, Negro male, and Negro female). The sample used in this portion of the study consisted of 20 students with I.Q.'s below 90 in each of the six race-sex groups for a total of one hundred twenty in the sample. All of the subjects were within three months of their six teenth birthday so that the norms of both tests could be applied. Each of the one hundred twenty subjects was administered the 22 subtests of the WISC and the WAIS in randomized order. The second hypothesis com pared intelligence (using the WISC and the WAIS) and reading achieve ment (using the Reading subtest of the Wide Range Achievement Test and the Gilmore Reading Test) as predictors of graduation in regular high school classes. In order to make comparisons between I.Q. and reading achievement possible, the reading achievement scores were converted into reading indexes which were ratios between actual grade placement and reading achievement level. The data for the second sample were obtained from clinical case study files of subjects who had in years past been tested as suspected mental retardates and for whom a . graduate or dropout status could be determined. The second sample consisted of 380 subjects, 183 of which had graduated successfully and 197 had dropped out. The data from both samples were tabulated and processed with electronic equipment. To evaluate hypothesis number one, data obtained were tested with the analysis of variance technique. The level of significance for all statistical tests was set at the .05 level. 144 Analysis of the comparisons between the three scales of the two tests for the several race and sex groups yielded the following results: 1. Overall WAIS-minus-WISC differences were highly signifi cant for Verbal, Performance, and Full Scale I.Q.’s. 2. Among the race groups, the Negroes had a significantly wider WISC-WAIS disparity than the Anglos or Mexicans. 3. There were no significant sex differences. Further inspection of the data revealed that: 1. In terms of the practical effects that the test differences would have on class placement decisions, in seventy-five per cent of the cases both tests would produce the same pLacement, but in twenty-five per cent of the cases the WISC would call for | EMR placement and the WAIS regular class placement. There was no case in which the reverse was true. ! 2. The Vocabulary, Comprehension, Similarities, and Block I Design subtests accounted for the largest portion of the ! i ' | Full Scale I.Q. difference between the two tests, but all I : ! ; j j of the subtest pairs made positive contributions except the j ! J ! i Coding-Digit Symbol pair which reversed the WAIS-surpassing- ] ' WISC trend. j | | To evaluate hypothesis number two, the data were tested by com- i iputing biserial correlation coefficients between I.Q. and graduate j i | status and between reading achievement and. graduate status. Then the | i significance of the difference between correlations was obtained. In addition the per cent of accuracy with which I.Q.'s and reading indexes 145 predicted success was computed and this value was compared to chance. Finally the predictive accuracy of the I.Q. was compared to that of the reading index. Analysis of these comparisons yielded the following results: 1. The Verbal I.Q. and graduate status correlation (.349) is significant beyond the .01 level, but the size of the corre lation is so small that it accounts for only 12 per cent of the variance, thus leaving 88 per cent of the variance unac counted for. 2. The Performance I.Q. and graduate status correlation (.031) indicates that there is no relation between Performance I.Q. and probability of graduating in regular high school classes for this population. i | 3. The Full Scale I.Q. and graduate status correlation (.223) | is significant beyond the .05 level, but the size of the cor- | relation is so small that it accounts for only five percent of the variance, thus leaving 95 per cent of the variance unac- I I counted for. j : i j 4. The reading index and graduate status correlation (.602) is significant beyond the .01 level and accounts for 36 per j cent of the variance, thus leaving 64 per cent unaccounted for. , j ' ! 5. Tests of the significance between correlations indicated J ] that the superiority of the reading index over the I.Q.'s as predictors of graduation in regular high school classes for this population was significant beyond the .01 level. j 146 6. Analysis of the practical effects of using the Full Scale I.Q. versus the reading index as predictors of successful high school graduation in regular classes revealed that Full Scale I.Q.'s predicted with fifty per cent accuracy (equal to chance) while the reading index predicted with 74 per cent accuracy (24 per cent better than chance). Within the limitations previously described the following conclu sions were considered tenable: 1. Below-average intelligence individuals obtain higher WAIS than WISC I.Q.'s. The discrepancy is greatest on the Verbal Scale, followed by the Full Scale, and least on the Performance Scale. i 2. The disparity is greater for Negro subjects than for j Mexicans or Anglos. ! ! 3. Three-fourths of the time the WISC and the WAIS recommend j ! the same class placement, but in one-fourth of the cases the j WISC recommends EMR placement while the WAIS recommends regu- ! i lar class placement. j i 4. The Vocabulary, Comprehension, Similarities, and Block ■ Design subtests account for the largest portion of the dif- j ferences between the WISC and the WAIS I.Q.'s, but alloof the i subtests make positive contributions except the Coding-Digit- i Symbol pair which reverses the WAIS-surpassing-WISC trend. I j 5. For all practical purposes the I.Q. is of little value as a predictor of successful graduation in regular classes for 147 suspected mental retardates, but the reading index does have considerable predictive value. The investigator recommends further research to explore other aspects of the relationship between these two I.Q. tests and their value as predictors of academic success. It is further recommended that other research be conducted in an effort to identify other factors which may account for success or failure in regular classes. APPENDIX A VERBAL, PERFORMANCE, AND FULL SCALE I.Q.'s AND SUBTEST SCALE SCORES (WAIS TRANSFORMED) FOR THE SUBJECTS IN SAMPLE 1 148 EXHIBIT 1 VERBAL, PERFORMANCE, AND FULL SCALE I.Q.'s AND SUBTEST SCALE SCORES (WAIS TRANSFORMED) W I S C FOR ANGLO MALE SUBJECTS IN SAMPLE 1 W A I S £ £ O •H to 0 + j £ D O 0 £ £ O •rl 0 w 0 p IH < D O ' IH o tn 0 •H X £ £ £ •H 0 0 O t/> •H £ o ' £ £ O 0 •H £ •rH p Pi a) O ' O 0 to ' £ rH •H £ •H 4 - > £ H O 0 rH P 0 + ■ > •H £ P i IH •H E ID 5 c d P 0 P •H P i £ IH •H E * 2 c d £ X C D Pi rH C /D Q)+J C D C DO rH £ 3 O £ ,£ 0 u C O rH 0 P 00 rH p O Pi CD G £ 3 rH Pi < D Pi bfl P .P D O C O E 0 E cd £ rH rH Pl0 £ Pi D O rH C O 0 Pi rP rH .O P C d £ £ O E £ . 0 H Pi rH p £ p 0 £H A S DOPE . ^ G O 0 X) O Pi +J •rH £ •H O P CLP £ T3 0 0 •H C H rH O Pi p •H •H £ ■H-P P & l 0 .H P E < D O M h rH o X E •H G 0 D O T5 P > E O Pi 0 •T H lO H rH cp E •H G D O C D Pi DOE OE O £ C D H •tH W ) fH iH £ Pi 0) C O Z £ 1 — 1 O U Pi < •rH C O 0 > •H a 0 > •H O P.CJ •H Pi £ . < rH C Q S 3 O CJ 0 O ' Pi PI P O ' C l* hH £ IH O u Pi < •H C O •H Q ’ O > 0 > o r a H O PCD rH 0*rlH CQQPlC < § 3 < 1 > a, P P H A M 0 1 0 6 0 8 0 7 1 0 0 7 1 0 0 8 7 0 9 1 0 0 9 0 7 0 9 0 9 2 0 8 8 0 7 0 6 0 7 1 1 0 9 0 8 0 9 0 1 0 0 7 0 9 0 7 0 8 0 9 0 0 8 9 A M 0 2 0 9 0 9 0 9 1 1 1 2 1 0 1 0 0 0 7 0 6 0 8 0 7 1 4 0 8 9 0 9 4 1 1 0 8 1 0 0 9 1 3 1 1 1 0 1 1 2 0 7 0 6 0 6 1 2 0 9 2 0 9 7 A M 0 3 0 2 1 0 0 2 0 3 0 5 0 6 0 6 6 0 4 0 5 0 4 0 5 0 6 0 6 4 0 6 2 0 3 0 6 0 3 0 4 0 5 0 8 0 7 1 0 7 0 7 0 5 0 7 0 8 0 7 9 0 7 3 A M 0 4 0 6 0 8 0 6 0 1 0 9 1 0 0 7 9 0 9 0 5 1 4 1 0 0 8 0 9 4 0 8 5 0 8 1 1 0 6 0 9 1 3 1 0 0 9 5 0 8 1 1 1 1 0 8 1 2 1 0 1 0 9 7 A M 0 5 0 6 0 7 0 6 0 6 0 8 0 7 0 7 9 0 6 0 7 0 8 0 9 1 0 0 8 6 0 8 0 0 8 1 0 1 4 1 0 0 8 0 7 0 9 7 0 9 1 1 12 1 0 1 4 1 0 6 1 0 1 A M 0 6 0 7 0 9 0 7 1 1 0 7 0 7 0 8 7 1 3 0 8 0 9 1 0 1 3 1 0 4 0 9 5 0 8 1 1 0 7 1 0 0 5 0 9 0 9 2 0 9 1 2 1 1 1 1 1 1 1 0 4 0 9 7 A M 0 7 0 7 0 7 0 6 0 5 0 6 0 4 0 7 5 1 1 0 8 0 9 10 1 3 1 0 1 0 8 6 0 8 0 6 0 7 1 4 0 6 0 7 0 9 1 12 1 3 1 1 0 6 1 0 1 0 2 0 9 5 A ¥ 0 8 0 5 0 7 0 8 0 8 0 6 1 1 0 8 5 1 1 0 5 1 2 10 11 0 9 9 0 9 1 0 4 0 8 0 6 0 6 1 3 0 7 0 8 7 0 9 0 9 1 1 0 7 0 9 0 9 1 0 8 8 A M 0 9 0 5 1 0 0 6 0 6 0 6 0 6 0 7 9 0 7 0 8 0 7 1 0 1 2 0 9 2 0 8 3 0 5 0 9 0 6 0 8 0 6 0 8 0 8 6 0 8 0 9 0 9 0 8 1 0 0 9 2 0 8 8 A M 1 0 0 6 0 0 0 6 1 0 0 8 0 6 0 7 5 0 8 1 3 1 0 1 2 1 0 1 0 4 0 8 8 1 0 0 8 0 6 1 0 0 6 1 0 0 9 2 0 8 1 2 1 2 0 9 1 2 1 0 4 0 9 7 A M 1 1 0 4 0 6 0 5 0 5 0 3 0 7 0 6 9 0 9 0 5 0 7 1 0 0 8 0 8 5 0 7 4 0 5 0 6 1 3 0 8 0 8 0 5 0 8 7 0 7 0 7 0 6 0 7 0 8 0 8 2 0 8 4 A M 1 2 0 9 1 4 0 5 0 8 1 2 0 7 0 9 5 1 1 0 8 0 7 1 0 0 3 0 8 5 0 8 9 . 11 1 8 0 6 1 1 0 6 1 4 1 0 7 0 7 1 2 0 9 0 6 1 2 0 9 4 1 0 2 A M 1 3 0 9 1 3 0 7 0 9 1 1 0 7 0 9 6 1 4 1 1 1 3 1 0 0 9 1 1 0 1 0 3 1 1 1 1 12 1 3 0 6 1 2 1 0 4 0 8 1 1 1 2 1 1 1 3 1 0 6 1 0 5 A M 1 4 0 9 0 7 1 3 0 8 0 7 0 9 0 9 2 1 3 1 2 0 9 12 1 3 1 1 3 1 0 2 0 9 0 9 1 0 0 8 1 1 0 7 0 9 5 1 1 1 2 1 1 0 8 1 3 1 0 4 0 9 9 A M 1 5 0 6 0 7 0 8 0 8 0 6 1 3 0 8 7 0 8 07. 1 1 13 0 7 0 9 4 0 9 0 0 9 0 7 0 9 0 8 0 9 0 8 0 9 3 0 7 0 8 1 1 0 7 0 8 0 9 1 0 9 1 A M 1 6 0 9 1 2 0 8 1 1 1 0 0 7 0 9 7 1 3 1 1 0 4 0 9 0 6 0 9 0 0 9 3 1 1 1 1 0 9 1 1 0 8 12 1 0 1 0 7 1 0 0 6 1 0 0 8 0 8 7 0 9 5 A M 1 7 1 3 1 2 0 7 1 2 1 2 0 9 1 0 8 0 9 1 1 0 7 1 6 1 0 1 0 4 1 0 7 0 8 1 3 0 5 1 1 0 9 1 1 0 9 6 0 8 0 9 0 8 10 1 1 0 9 4 0 9 5 A M 1 8 1 0 0 9 0 8 1 1 1 3 0 6 0 9 7 1 3 0 9 1 1 1 0 1 4 1 1 0 1 0 4 1 0 0 9 1 1 1 6 0 8 1 3 1 0 6 1 2 0 8 12 1 1 1 3 1 0 7 1 0 7 A M 19 0 5 0 9 0 8 0 8 0 6 0 4 0 7 9 1 3 0 9 0 8 1 3 15 1 1 1 0 9 3 0 5 0 9 1 2 0 9 0 6 0 7 0 9 2 1 3 0 9 12 0 5 0 9 0 9 6 0 9 3 A M 2 0 0 5 0 5 0 6 0 8 0 4 0 6 0 7 2 0 9 0 8 1 0 0 6 1 5 0 9 7 0 8 3 0 5 0 6 0 6 0 6 0 3 0 6 0 7 4 0 9 0 9 1 0 0 7 1 2 0 9 9 0 8 4 EXHIBIT 1 (Continued) VERBAL, PERFORMANCE, AND FULL SCALE I.Q.'s AND SUBTEST SCALE SCORES (WAIS TRANSFORMED) FOR ANGLO FEMALE SUBJECTS IN SAMPLE 1 W I S C W A I S O ' IH 0 iH € O • rH G O •H to fj o •rH t f l 0 •H 4J <3 o' G O -p G 0 G DO •H to ‘ 0 0 G 0 > — 4 G O •rH g •H in G 0 •rH in 0 •rH +J S O ' G O G bO •H cn ‘ +-> G 0 0 0 § «L) CD 4J •rH cd P, 4 —4 •rH £3 0 cd cd •P 0 H-> •rH G. cd 4-4 •rH 0 £ 1 e cd p CD cd •G CD p CD 6 r- u cd H i — i G X) CO 4-> t — I cd 0 -p P 0 Gt-i 0 0 fH DC 3 C Q 1 OS rH tJi bo G . O O CO cd 0 P 0 s .G ?H cd rH CO •P rH G •9 rH cd 1 —4 -4-0 04-4 P 0 GrH a r* CD 0 1 —1 p bOPXi 3 R P 6 £ P 18 O CO CD ,0 o p, 4J •rH cd •rH ,o 4-4 Q.4-4 cd O <D<D •rH C+H 1 —4 O G, H-> •H •rH cd •p 0 4-4 to HH 1 — 1 O cd OS X CD CO j C+H G 1-4 s O u •H P < E •rH CO o o > DO •H a ' P 0 > u g •H O 0 , 0 O P O 1 — ( CO •Pin T3 O CJ P 0 O ' 0,4-4 t —4 G O ' CO 1 — 4 tp G 4-4 E O U •rH P < E •rH CO bO •rH Q * O O > p 0 > C M • H O P l U 0 r —4 cn U MT-Jin •H H^acn a, < o < ?H o c y O.HH , — 4 G CP • A F 01 11 12 08 11 13 09 104 10 10 02 07 12 087 096 12 11 09 13 11 13 107 11 13 08 10 06 095 102 A F 02 08 11 06 11 07 09 091 08 08 08 09 10 090 090 10 10 10 13 09 09 101 11 12 10 06 08 092 097 A F 03 05 04 04 06 03 04 070 07 07 05 06 09 078 071 04 06 05 07 03 07 073 10 05 06.:06 08 081 075 A F 04 08 10 13 09 06 11 097 08 09 08 08 15 097 097 10 11 10 10 11 08 100 13 11 09 11 10 104 102 A F 05 09 09 07 09 07 09 090 05 05 08 10 11 085 086 09 10 08 11 11 08 097 10 06 06 07 10 087 093 A F 06 06 05 04 06 05 07 072 09 07 08 09 11 094 081 04 06 06 05 08 06 076 12 08 10 07 09 094 083 A F 07 04 05 05 02 04 07 066 04 09 01 10 11 079 070 04 06 05 00 08 04 071 10 08 04 07 08 083 075 A F 08 08 08 07 08 07 10 087 11 09 10 07 08 093 089 11 14 06 13 09 11 104 07 10 11 11 06 095 100 A F 09 07 07 08 11 06 10 089 11 08 13 14 12 111 099 09 11 08 10 11 09 104 11 11 12 12 12 108 107 A F 10 06 05 08 08 04 02 072 07 09 08 10 12 094 081 06 09 06 09 05 07 085 10 10 11 06 10 092 087 A F 11 08 07 07 05 08 09 084 05 01 04 05 12 068 074 10 08 06 11 08 08 093 12 09 08 06 03 081 087 A F 12 09 05 08 08 08 07 085 10 11 11 10 13 107 095 09 07 10 08 12 09 092 13 10 13 12 14 114 101 A F 13 06 05 13 12 08 07 092 10 11 08 06 12 096 093 08 09 09 10 08 07 090 11 10 09 06 10 094 091 A F 14 06 05 04 05 06 09 074 05 04 05 08 10 075 072 05 09 06 09 08 08 089 09 04 06 07 08 081 084 A F 15 07 03 04 07 08 07 075 08 09 07 10 07 087 079 09 12 05 09 08 11 094 10 08 06 06 06 079 087 A F 16 02 04 03 03 02 06 058 06 06 08 10 13 090 071 03 06 06 06 06 02 070 11 06 08 06 08 085 075 A F 17 03 04 06 06 02 06 066 10 05 08 07 11 087 074 03 03 05 06 06 04 072 09 06 06 09 08 085 076 A F 18 04 05 04 05 05 07 069 06 05 05 09 13 083 073 04 07 06 08 11 06 085 13 08 08 06 11 092 087 A F 19 09 06 08 07 06 06 081 06 06 07 10 12 087 083 08 09 09 10 08 08 093 09 07 10 11 09 094 093 A F 20 04 03 07 04 06 04 066 04 01 04 05 05 057 058 10 09 06 08 05 07 086 06 06 09 04 08 078 081 cn o EXHIBIT 2 W I S C p 0 0 ) r O o X E o J 0 3 es co :z ; M M 01 M M 02 M M 03 M M 04 M M 05 M M 06 M M 07 M M 08 M M 09 M M 10 M M 11 M M 12 M M 13 M M 14 M M 15 M M 16 M M 17 M M 18 M M 19 M M 20 r o •H P a s o <P R 1 —I r o •rl r © ,R © P s' o u w © •H P •H P a S I — ! •H E •rl CO 06 06 06 05 06 07 04 07 10 12 06 07 03 05 09 11 06 08 10 05 06 05 10 07 10 09 06 05 10 05 05 06 06 07 04 06 06 05 04 06 o •rl P 0 ) E •R P •H s 09 10 07 03 04 04 07 05 09 11 05 03 02 06 08 11 07 05 07 11 06 08 11 11 09 11 04 08 08 09 08 09 02 04 05 03 04 08 05 08 x P a S t — I 3 3 o o > s a CO p • r H b O •rl Q ' R O •rl P 0 i— l i* O O' u 1 - 1 0 p 3 P o •rl CP VERBAL, PERFORMANCE, AND FULL SCALE I.Q.'s AND SUBTEST SCALE SCORES (MIS TRANSFORMED) MEXICAN MALE SUBJECTS IN SAMPLE 1 W A I S ■ 8 0 > R P t» 0 R *rl 0 t n E 0 0 0 a WM Pg.rX pas o up o •rp P rQ bo O E R 0 0 -H •n lflT l X> W O O' 0 O § E P o cp P 0 a, O' l —l 0 rH O S O CO rH 3 CP 08 07 06007 02 04 04 10 07 04 06 07 00 09 07 10 07 10 10 04 05 09 08 07 09 11 05 07 06 07 07 09 06 06 00 02 04 07 00 06 CQ 0 < CJ 085 07 09 01 10 11 083 083 072 06 08 09 10 10 090 079 066 10 08 09 05 06 083 072 076 04 01 10 08 10 076 074 092 10 09 13 15 09 108 100 072 11 06 09 06 06 083 075 063 07 10 10 08 12 096 077 096 11 10 15 15 11 115 106 082 10 09 08 10 12 099 089 086 09 05 08 06 07 079 081 079 07 11 08 08 07 087 081 094 18 11 14 15 10 125 109 099 09 06 08 05 08 080 089 074 09 04 09 12 10 092 080 085 09 08 10 11 10 097 090 084 08 12 14 15 15 120 101 076 05 11 07 11 11 093 083 060 05 04 05 06 08 069 061 080 07 01 02 06 10 067 071 067 05 01 02 05 05 055 058 R O •rl P as O t p R R O •rl tn R 0 ,R 0 P f CJ o •rl P 0 P •rl tn 0 •rl P •rl P a S rH •rl E •rl CO 5 6 P •rl bO •rl O' X rH 3 ■3 u o > O' a S -P 0 > o rQ CO p •rl b O •rl CT R O •rl P 0 rH i* o CJ 0 p 3 P o •rt a, R •rl OT* 0 Q O O P R 0 E 0 0 I — I p bOP 3 R .O S PTO © © O Pvt C D O' l —l 0 O g O' I — I 0 t — 1 a S o CO 07 11 08 10 06 09 093 10 13 08 09 08 08 08 07 089 07 07 08 06 05 08 06 06 082 07 07 02 06 07 08 09 06 082 09 04 09 10 09 06 08 09 092 07 08 04 09 05 06 08 05 079 06 11 04 06 06 04 06 03 074 11 12 11 07 09 08 08 07 092 09 13 09 09 07 05 11 07 089 13 11 11 09 06 08 05 11 092 05 04 08 06 06 05 06 07 080 08 08 11 12 09 11 11 08 102 11 18 10 10 10 08 12 09 100 07 09 08 09 06 11 08 06 089 10 12 08 11 12 13 09 07 100 09 08 05 10 10 09 06 06 089 12 10 08 06 06 06 06 07 081 11 07 05 05 06 07 05 05 076 07 04 03 06 03 06 08 05 074 08 06 04 03 06 07 11 05 079 07 04 IfKfl 1 0 oq eu< o < 09 09 15 06 07 10 10 07 11 12 06 13 11 09 06 12 07 08 11 06 09 12 09 11 10 07 13 09 08 07 06 07 10 12 09 08 08 12 08 10 06 14 13 08 12 13 07 11 05 06 10 05 08 03 06 06 08 06 07 03 3 C P o tp p 0 CP 108 099 085 086 089 084 094 086 087 089 092 084 095 082 102 096 102 094 082 087 086 081 107 105 091 096 103 094 103 102 103 094 083 080 074 073 078 074 074 075 cn EXHIBIT 2 (continued) VERBAL, PERFORMANCE, AND FULL SCALE I.Q.’s WISC -AND SUBTEST SCALE SCORES (WAIS TRANSFORMED) WAIS MEXICAN FEMALE SUBJECTS IN SAMPLE 1 £ g . ' ' 0) u 0 ,n £ o •H cd o I- o •rH < / ) £ 0 .£ 0 £ P. o •rl + - > 0 e x +J < D •rH P •H P cd rH •rH & cd rH 3 •3 £ Ph co + - > •H O' i —i i —i cd ■p £ £ O 0 •H E 0+J 00 £ 0 fnbo PrH 3£ +j pj->cd £ bo •H t / > ‘ 0 n A! CJ i -i + J -S O S 0 0 bO £. •M 0 o § 0 rH cd o CO rH £ o •H cd o o •H t n £ 0 ,£ 0 £ & o •rl + - > 0 + - > to 0 •H P •rH u cd rH •H u CO ■p •H & cd i —i 3 id O' i —i rH cd •f rH O f CO + ■ > •rl £ bO £ -H. O V) •rl 0 0 + - > Q £ 0 3rl A! P f t U :ture •angement ect embly •formance t-H 0 . —! cd CJ CO rH o x m E £ o bO p O S O M O nd £ rH 4h E •rl £ bO 0 fH bO o 0 H T"*tn H i —1 Cd oz o co J £ i —i O u <: •H CO o > •rH O' 0 > •rl 0»r(Jil fXCJQ. i —1 03 ni/ 1 O o < u 0 O' PL . I -I 3 O' £. 1 —1 £ i —i O u •H cn •H Q o > 0 > •rl O' •rl O rH C L . CJ O Q •HHpm t u cr cl,< o<eu-i 3 £ . M F 01 04 05 08 09 06 10 082 13 05 08 08 10 092 085 08 09 06 09 09 07 091 09 09 10 09 08 091 090 M F 02 02 05 03 07 02 04 061 04 06 06 08 10 078 066 04 06 03 09 05 06 077 09 11 06 04 09 083 078 M F 03 05 05 05 07 05 04 070 04 05 05 11 06 074 069 04 10 06 09 08 08 088 09 09 06 07 11 089 087 M F 04 07 07 08 11 06 09 087 10 08 05 03 12 083 084 09 09 07 10 08 10 094 09 03 08 08 08 083 088 M F 05 005 07 06 08 05 06 076 10 07 09 16 09 101 087 03 09 06 08 06 06 082 09 06 10 06 10 087 083 M F 06 05 06 03 11 05 06 075 07 11 05 10 10 090 080 05 06 06 09 06 05 081 07 09 08 11 09 090 084 M F 07 04 06 05 07 04 09 074 06 06 08 09 10 085 077 04 06 06 10 08 07 083 11 07 06 07 09 085 083 M F 08 05 06 04 04 05 09 072 11 09 10 10 12 103 085 07 07 05 06 06 06 078 09 11 12 08 12 103 088 M F 09 08 07 03 08 05 07 077 05 05 07 05 13 079 076 07 07 05 07 09 07 083 09 06 06 06 06 077 079 M F 10 06 07 04 09 06 07 079 06 06 10 10 10 089 082 07 09 02 08 08 07 084 09 08 10 07 09 090 086 M F 11 03 06 04 05 05 04 066 09 10 07 09 10 093 077 03 06 03 05 05 07 072 09 07 08 07 09 089 078 M F 12 04 07 06 08 04 07 075 08 08 09 07 13 093 082 04 06 03 08 08 07 078 12 09 06 09 08 091 083 M F 13 06 05 07 09 04 07 077 07 05 08 03 09 075 074 04 07 05 09 08 07 082 09 07 09 08 03 085 082 M F 14 05 07 06 08 06 09 080 05 08 02 09 11 079 077 04 06 06 04 09 08 080 10 09 09 09 13 100 088 M F 15 09 07 07 08 05 06 081 05 08 01 09 09 075 076 06 09 07 13 08 07 091 09 09 07 08 10 094 091 M F 16 07 07 05 08 07 07 080 08 04 02 06 13 076 076 08 07 06 06 08 07 084 07 09 06 06 09 081 081 M F 17 06 05 05 08 04 06 077 06 06 05 10 12 085 079 08 08 05 08 06 06 082 09 07 06 06 10 085 082 M F 18 05 08 05 06 08 07 079 08 07 07 12 07 087 081 06 08 06 06 06 07 081 08 08 06 06 16 092 085 M F 19 06 06 08 10 05 07 081 11 12 11 10 11 107 093 05 06 05 08 08 07 082 11 11 11 08 15 107 092 M F 20 05 07 03 06 05 04 069 09 04 08 10 11 089 076 06 06 07 07 08 06 081 10 07 08 06 13 092 085 W I S C EXHIBIT 3 VERBAL, PERFORMANCE, AND FULL SCALE I.Q.’ s AND SUBTEST SCALE SCORES (WAIS TRANSFORMED) NEGRO MALE SUBJECTS IN SAMPLE 1 W A I S p C D C D c d < d 3 05 C/3 2 G O •H P cd < 8 c 0 5 o •H tn 0 5 C D 4 3 C D P I* O PH CJ o •H P C D 6 43 P •H O h C rH O' ,g H H O' tn G B IH C D P b O C D C D G • r H G G G-H. tn C J C D O P 0? cd O' O cd tn tn G rH •H •H G Oc IH •H S < D < cd cd P O h * — i C O C D P C D C D C3 C J G cd 3 rH O h < D O h b O p b O H C O B rH 43 P cd P H 3 G 44 O G. O p •H cd • r H 43 P CUP G o C D •H C p iH O B cj b O O h OBUOhO •r-i r3 O h rH cp • r H o •rH (D •H O.H O h rH 43 O C D 3 G C O > Q ' > &, CJ C L , < £Q O CJ Oh P-, IH 13 O • H tn G C D 43 C D O h w O u O' p C D £ 43 P •rH w C D •H P •rH Oh cd rH •rH £ •H C/J G cd Ph CO P •H bO •H Q ' f c * G f “ 4 P • s C J o > O' cd * 8 C D > o H CO 6 > G o tn •H C D C D P Q O h (D P 3 < -H 44 •rt P P« C J bd C J B o •H » pH rQ OP,° < D C J X l P G C D £ C D C D . . „ O h b 0 P , n O h 3 G P B O P cd C D C D C J O h •!->(/) O h ■h O h , o tn C D ' 0 < P . C Q Cu< O' C D r-H cd o CO 3 N M 01 08 06 06 09 09 06 084 16 10 10 10 09'7107 h 094 09 09 06 11 09 08 094 11 12 11 07 11 100 096 N M 02 06 06 04 06 05 07 072 04 05 06 04 09 069 068 04 06 05 07 08 07 079 09 07 08 07 09 086 081 N M 03 10 11 04 09 09 07 090 11 10 05 11 13 100 094 10 14 07 11 08 12 101 11 11 11 09 12 106 103 N M 04 05 09 06 09 05 09 082 06 04 05 06 07 069 074 06 07 07 08 09 06 086 09 03 05 08 08 078 081 N M 05 04 05 04 06 02 09 069 10 09 11 12 10 103 083 05 06 06 07 06 05 079 09 11 11 07 12 099 087 N M 06 08 09 04 05 06 04 075 06 05 03 06 03 062 066 06 07 06 08 06 08 083 04 07 06 06 09 078 080 N M 07 09 11 11 13 10 13 108 07 07 05 06 09 078 093 11 13 10 11 12 10 104 10 09 03 09 09 085 096 N M 08 06 09 05 07 06 09 081 11 07 05 11 08 089 083 07 06 05 09 08 08 084 06 13 09 10 08 094 087 N M 09 08 08 08 08 07 09 087 07 08 05 08 07 079 082 08 09 08 09 06 10 092 07 06 08 06 10 082 087 N M 10 05 09 05 04 02 09 072 06 07 05 06 06 072 070 04 08 06 06 09 05 081 05 05 03 07 05 070 075 N M 11 04 06 03 06 02 04 069 07 10 08 10 08 090 077 04 09 06 06 11 06 084 09 06 08 07 08 082 082 N M 12 05 07 05 06 05 04 071 07 10 08 10 10 093 080 05 07 07 08 05 07 082 08 07 08 06 09 085 082 N M 13 08 07 08 08 05 15 091 08 07. 07 08 07 082 085 08 12 08 11 12 08 096 12 04 06 06 05 079 088 N M 14 05 07 06 06 04 06 072 04 07 05 02 07 065 066 04 07 06 08 06 06 081 07 07 05 06 04 073 076 N M 15 06 07 03 08 07 06 076 10 06 02 14 12 092 082 08 10 06 08 09 08 091 12 10 07 07 16 103 096 N M 16 04 11 06 08 05 04 077 11 07 11 12 09 100 087 07 11 05 09 06 10 089 09 05 11 06 04 083 086 N M 17 06 04 06 01 05 04 065 10 11 05 12 09 096 077 05 08 07 08 06 08 086 08 12 08 06 12 094 098 N M 18 07 04 04 09 06 09 079 11 08 12 14 06 101 088 07 09 07 12 09 08 094 06 15 09 10 13 104 099 N M 19 05 05 04 06 07 06 072 10 01 05 12 07 079 073 04 08 07 07 08 07 083 09 11 08 09 11 098 088 N M 20 06 10 05 10 07 14 091 13 09 10 09 12 104 097 04 09 07 08 17 09 097 09 09 06 07 06 083 091 M c n oa ( EXHIBIT 3 (Continued) VERBAL, PERFORMANCE, AND FULL SCALE I.Q.'s W I' S C AND SUBTEST SCALE SCORES (WAIS TRANSFORMED) WAIS NEGRO FEMALE SUBJECTS IN SAMPLE 1 5 « o o •H *H • • p o' 4 J o' 0 O' a > r-i • o' f H 0 0 JD o x e r t o 3 rt co 2: N F 01 N F 02 N F 03 N F 04 N F 05 N F 06 N F 07 N F 08 N F 09 N F 10 N F 11 N F 12 N F 13 N F 14 N F 15 N F 16 N F 17 r t Q C O r r t w o u C D f H r i r * w cn 0 ) • r t r r t & • r t f •a M • c o *H • r t C O o • r t o • r t jj X t j § O' 4 - > A H bO • r t t n * C D O H < D s o U • r t c n r t o c c d ft CO t O' r r t o ■§ W O ■ p c E O cn 0 O I r t C D 4 - > r t H o r * P C D t - • r t J j n rt ft CO p M s C D 0 f H W B U O f H , _ J ( j V i C D fs r t o r t 0 \ • r H ■ p c d g o • r t ■ p 0 P • r t t. r t M £ CJ cn * C D < D s cn < r r t r t rH o M n r t r t r \ f r t r t 1 o o Obj ect Assemb] bo e * u cn r r t < D t. r t _ i r r t 3 r r t r t • 8 0 > C / D C D U H Q 0 0 f H &0 3 f 3 •P c d O f H P o cn o < + H c & Q rH P • r t t j • r t £ • r t r t o Q • r t bO • r t • 8 C D > • P o ■H h . • r t n u U i r t f r t r * H o cn H H ■p •H e . pi • r t s P • r t bO id o P • r t bO r t • P U O O o ' C D • i r t 4h f H r r t r r t H w u CO > Q* ft ft <2 cq u CJ u / Oh tu > H I r t o u % • r t CO • r t Q* o > • r t Q* •H ft i r t CO ft<! rO o 0 ft p S 06 09 06 09 05 04 079 08 08 04 06 06 075 075. 07 09 06 10 05 07 085 07 09 05 08 06 081 082 08 10 09 13 07 11 097 10 08 08 10 17 104 101 09 14 08 13 11 10 104 12 08 09 08 12 100 103 05 04 03 01 05 07 063 05 06 05 11 11 083 070 08 02 06 06 06 08 079 09 07 06 10 14 094 085 07 07 05 06 10 04 079 05 08 09 09 09 086 080 13 09 06 08 05 12 094 08 07 08 13 12 098 095 04 04 03 03 04 04 060 05 04 02 00 04 051 051 04 05 02 04 08 07 072 06 07 03 07 04 073 071 05 06 06 08 05 13 082 05 03 07 05 11 074 076 04 07 10 07 12 07 089 09 05 06 05 10 081 084 05 06 06 06 04 07 072 08 05 08 06 12 085 076 07 07 05 05 11 07 082 10 07 09 09 06 089 084 06 06 08 08 06 07 080 07 05 07 08 12 085 080 06 08 08 10 08 08 089 10 08 08 10 14 099 093 08 08 08 10 08 06 087 05 08 07 08 12 086 085 10 11 10 13 08 09 100 12 09 09 11 10 103 102 06 06 08 09 06 11 085 09 12 13 14 15 118 101 08 07 07 08 09 08 089 11 07 11 07 10 095 091 06 10 06 06 08 15 091 09 03 06 07 13 083 086 09 07 07 11 15 12 101 12 07 08 06 12 094 098 04 04 06 07 05 09 074 06 05 07 05 11 078 073 04 06 10 04 11 05 081 09 10 06 05 09 083 081 09 07 06 08 07 10 086 13 12 09 06 16 108 096 08 08 06 13 09 08 094 17 11 06 09 08 100 096 04 05 06 08 05 13 080 05 01 08 05 10 071 073 04 06 07 08 11 06 085 09 07 08 04 08 082 083 06 09 07 05 06 06 079 07 09 04 08 13 087 081 07 08 07 08 06 07 085 11 07 03 10 06 083 083 05 09 06 06 05 07 077 07 08 02 03 10 072 072 08 09 07 09 11 06 092 09 09 03 06 07 075 084 06 10 07 10 06 11 090 11 06 09 06 16 097 093 08 12 08 11 13 10 102 14 09 09 11 10 104 103 M tn ■ r * APPENDIX B ANALYSIS OF VARIANCE OF SAMPLE 1 WISC-WAIS I.Q. DIFFERENCES VERBAL, PERFORMANCE AND FULL SCALE 155 156 EXHIBIT 1 ANALYSIS OF VARIANCE ; VERBAL I.Q.'s MEAN SOURCE SS DF SQUARE (a) Differences among Sample Means Race 1,233.0250 2 616.512 Sex 301.5042 1 301.504 Race x Sex 500.5583 2 250.279 Total for Sample Means 2,035.0875 5 (b) Overall Differences Between Tests 4,059.0375 1 4,059.038 (c) Interaction Between Tests and Sample Means Tests . Race 42.1750 2 21.088 Tests . Sex 2.6041 1 2.604 Tests . Race . Sex 20.0084 2 10.004 Total for Cell Means 6,158.9125 11 ERROR TERMS (d) Variation of Individuals Within Samples. 19,090,8750 114 167.464 (e) Pooled Subject .Test In- teractions within Samples 1,952.6750 114 17.129 21,043.5500 228 TOTALS (f) Total SS, all Observations Cell Means 6,158.9125 11 Variations within Cells 21,045.5500 228 27,202.4625 239 * F Significant at the .05 level but not at the .01 level *** F Significant beyond the .001 level. F 3.685* 1,802 1.497 236.969*** 1.231 0.152 0.584 157 EXHIBIT 2 ANALYSIS OF VARIANCE : PERFORMANCE I.Q.'s MEAN SOURCE SS DF SQUARE (a) Differences among Sample Means Race Sex Race x Sex 1,131.1000 660.0167 508.4333 2 1 2 565.550 660.017 254.217 Total for Sample Means 2,299.5500 5 - 00 Overall Differences Between Tests 380.0167 1 380.017 (C) Interactions Between Tests and Sample Means Tests . Race Tests . Sex Tests . Race . Sex 28.9333 109.3500 21.7000 2 1 2 14.467 109.350 10.850 Total for Cell Means 2,839.5500 11 - ERROR TERMS: (d) Variation of Individuals Within Samples 27,001.2715 114 236.853 Ce) Pooled Subject. Test In teractions within Samples 5,301.0285 114 46.5002 Total SS within Cells 32,302.500 228 - TOTALS (£) Total SS, all Observations Cell Means Variations within Cells 2,839.5500 32,302.3000 35,141.8500 11 228 239 - 2 2 1 * F significant at the .05 level, but not at the .01 level. ** F significant at the .01 level, but not at the .001 level. F .388* .787 .073 .172** .311 .352 .233 EXHIBIT 3 ANALYSIS OF VARIANCE : FULL SCALE I.Q.’s MEAN SOURCE SS DF SQUARE F (a) Differences among Sample Means Race Sex Race . Sex 1,088.2333 576.6000 577.6000 2 1 _2_ 544.117 576.600 288.800 3.236* 3.429 1.717 Total for Sample Means 2,242.4333 5 - - (b) Overall Difference Between Tests 2,470.4167 1 2470.417 112.998*** (c) Interactions between Tests and Sample Means Tests . Race Tests . Sex Tests . Race . Sex 35.4333 46.8167 16.0333 2 1 2 17.717 46.817 8.017 0.810 2.141 0.367 Total for Cell Means 4,811.1333 11 - - ERROR TERMS; (d) Variations of Subjects Within Samples 19,171.5000 114 168.171 - (e) Pooled Subject. . Test Inter * actions within Sample 2,492.3000 21,663.8000 114 228 21.862 TOTALS Cf) Total SS, all Observations Cell Means Variation within Cells 4,811.1333 21,663.8000 26,474.9333 11 228 239 - - * F significant at the .05 level, but not at *** F significant beyond the .001 level. the .01 level. APPENDIX C WISC AND WAIS I.Q.'s FOR SAMPLE 1 MEANS, STANDARD DEVIATIONS AND INTERCORRELATIONS 159 160 EXHIBIT 1 POPULATION NUMBER 1 : 120 CASES Variable Means Standard Deviation WISC Verbal I.Q. 80.11 10.76 Performance I.Q. 88.51 13.96 Full Scale I.Q. 82.52 11.17 WAIS Verbal I.Q. 87.83 9.49 Performance I.Q. 90.95 9.98 Full Scales I.Q. 88.80 8.61 Correlation Coefficients WISC IQ Test WAIS IQ Test Verb Perf F.Scale Verb Perf F.Sca! WISC Verbal 1.000 0.388 0.795 0.755 0.358 0.715 Performance 1.000 0.864 0.324 0.726 0.636 Full Scale 1.000 0.620 0.666 0.802 WAIS Verbal 1.000 0.422 0.833 Performance 1.000 0.801 Full Scale 1.000 161 EXHIBIT 2 ALL RACES MALE Variable Means Standard Deviation WISC Verbal I.Q. 81.41 11.45 Performance I.Q. 90.85 15.10 Full Scale I.Q. 84.46 11.79 WAIS Verbal I.Q. 88.43 9.62 Performance I.Q. 92.01 10.35 Full Scale I.Q. 89.95 8.722 Correlation Coefficients WISC I.Q.TEST WAIS Verb Perf F.Scale Verb WISC Verbal 1.000 0.363 0,778 0.667 Performance 1.000 0.863 0.362 Full Scale 1.000 0.593 WAIS Verbal 1.000 Performance Full Scale Perf F.Scale 0.270 0.710 0.615 0.455 1.000 0.619 0.680 0.783 0.789 0.829 1.000 ALL RACES FEMALE Means Standard Deviation WISC Verbal I.Q. 78.81 9.94 Performance I.Q. 86.18 12.40 Full Scale I.Q. 80.58 10.24 WAIS Verbal I.Q. 87.23 9.40 Performance I.Q. 89.88 9.56 Full Scale I.Q. 87.66 8.42 Correlation Coefficients WISC I.Q. TEST WAIS I.Q. TEST Verb Perf F.Scale Verb Perf F.Sca! WISC Verbal 1.000 0.393 0.809 0.861 0.448 0.822 Performance 1.000 0.857 0.267 0.742 0.567 Full Scale 1.000 0.652 0.719 0.818 WAIS Verbal I D . 000 0.379 0.882 Performance 1.000 0.765 Full Scale 1.000 162 EXHIBIT 3 ANGLO BOTH SEXES Variable Mean Standard Deviation WISC Verbal I.Q. 82.55 12.52 Performance I.Q. 91.72 12.71 Rill Scale I.Q. 85.55 11.51 WAIS Verbal I.Q. 90.47 10.96 Performance I.Q. 93.12 9.76 Full Scale I.Q. 91.50 9.28 Correlation Coefficients WISC I.Q. TEST WAIS I.Q. TEST Verb Perf B.Scale Verb Perf F.Sca! WISC Verbal 1.000 0.421 0.838 0.770 0.421 0.764 Performance 1.000 0.842 0.350 0.799 0.627 Full Scale 1.000 0.663 0.725 0.828 WAIS Verbal 1.000 0.454 0,887 Performance 1.000 0,773 Full Scale 1.000 MEXICAN BOTH SEXES Mean Standard Deviation WISC Verbal I.Q. 77.65 8.94 Performance I.Q. 88.25 14.23 Full Scale I.Q. 81.02 10.61 WAIS Verbal I.Q. 83.92 8.05 Performance I.Q. 90.80 9.15 Full Scale I.Q. 86.35 7.26 Correlation Coefficients WISC I.Q. TEST WAIS I.Q. TEST Verb Perf F.Scale Verb Perf F.Sca] WISC Verbal 1.000 0.415 0.774 0.555 0.375 0.695 Performance 1.000 0.896 0.185 0.666 0.607 Full Scale 1.000 0.394 0.639 0.755 WAIS Verbal 1.000 0.406 0.731 Performance 1.000 0.830 Full Scale 1.000 EXHIBIT 3 (continued) NEGRO BOTH SEXES Mean Standard Deviation WISC Verbal I.Q. Performance I.Q. Full Scale I.Q. WAIS Verbal I.Q. Performance I.Q. Full Scale I.Q. WISC Verbal Performance Full Scale WAIS Verbal Performance Full Scale 80.15 10.20 85.57 14.53 81.00 11.01 89.10 8.09 88.92- 10.77 88.57 8.59 Correlation Coefficients WISC I.Q. TEST Verb Perf F.Scale 1.000 0.333 0.764 1.000 0.861 1.000 WAIS I.Q. TEST Verb Perf F.Scale 0.888 0.267 0.632 0.470 0.701 0.692 0.795 0.615 0.805 1.000 0.456 0.835 1.000 0.844 1.000 164 EXHIBIT 4 ANGLO MALE WISC Verbal I.Q. Performance I.Q. Full Scale I.Q. WAIS Verbal I.Q. Performance I.Q. Full Scale I.Q. Mean Standard Deviation 85.70 12.23 96.20 11.75 89.50 10.61 91.85 10.26 96.05 8.19 93.85 7.99 Correlation Coefficients WISC I.Q. TEST WAIS I.Q. TEST Verb Perf F.Scale Verb Perf F.Scale WISC Verbal 1.000 0.335 0.801 0.656 0.167 0.605 Performance 1.000 0.821 0.378 0.711 0.605 Full Scale 1.000 0.616 0.541 0.737 WAIS Verbal 1.000 0.457 0.856 Performance . 1.000 0.770 Full Scale 1.000 ANGLO FEMALE Mean Standard Deviation WISC Verbal I.Q. 79.40 12.30 Performance I.Q. 87.25 12.30 Full Scale I.Q. 81.60 11.25 WAIS Verbal I.Q. 89.10 11.71 Performance I.Q. 90.20 10.51 Full Scale I.Q. 89.15 10.06 Correlation Coefficients WISC I.Q. TEST WAIS I.Q. TEST Verb Perf F.Scale Verb Perf F.Scale WISC Verbal 1.000 0.394 0.851 0.870 0.537 0.869 Performance 1.000 0.817 0.288 0.831 0.589 Full Scale 1.000 0.707 0.810 0.879 WAIS Verbal 1.000 0.430 0.917 Performance 1.000 0.745 Full Scale 1.000 165 EXHIBIT 4 MEXICAN MALE Mean Standard Deviation WISC Verbal I.Q. Performance I.Q. Full Scale I.Q. WAIS Verbal I.Q. Performance I.Q. Full Scale I.Q. 79.40 89.85 82.95 85.15 91.90 88.05 11.01 17.80 13.62 10.17 10.76 9.24 Correlation Coefficients WISC I.Q. TEST WAIS I.Q. TEST Verb Perf F. Scale Verb Perf F.Scal WISC Verbal 1.000 0.522 0.816 0.504 0.460 0.708 Performance 1.000 0.918 0.307 0.698 0.687 Full Scale 1.000 0.435 0.682 0.790 WAIS Verbal 1.000 0.589 0.757 Performance 1.000 0.884 Full Scale 1.000 MEXICAN FEMALE Mean Standard Deviation WISC Verbal I.Q. 75.90 6.03 Performance I.Q. 86.65 9.66 Full Scale I.Q. 79.10 6.16 WAIS Verbal I.Q. 82.70 5.16 Perfoimance I.Q. 89.70 7.31 Full Scale I.Q. 84.65 4.10 Correlation Coefficients WISC I.Q. TEST WAIS I.Q. TEST Verb Perf F.Scale Verb Perf F.Scale WISC Verbal 1.000 0.002 0.573 0.678 0.100 0.578 Performance 1.000 0.819 -0.320 0.574 0.257 Full Scale 1.000 0.121 0.509 0.523 WAIS Verbal 1.000 -0.155 0.580 Performance 1.000 0.711 Full Scale 1.000 166 EXHIBIT 4 (Continued) NEGRO MALE Variable Mean Standard Deviation WISC Verbal I.Q. 79.15 10.35 Performance I.Q. 86.50 14.22 Full Scale I.Q. 80.95 9.55 WAIS Verbal I.Q. 88.30 7.44 Performance I.Q. 88.10 10.79 Full Scale I.Q. 87.95 7.95 Correlation Coefficients WISC I.Q. TEST WAIS I.Q. Verb Perf F.Scale Verb Perf WISC Verbal 1.000 0.053 0.662 0.866 0.003 Performance 1.000 0.782 0.370 0.663 Full Scale 1.000 0.815 0.487 WAIS Verbal 1.000 0.280 Performance 1.000 Full Scale NEGRO FEMALE Mean Standard Deviation WISC Verbal I.Q. 81.15 10.22 Performance I.Q. 84.65 15.14 Full Scale I.Q. 81.05 12.55 WAIS Verbal I.Q. 89.90 8.81 Performance I.Q. 89.75 10.95 Full Scale I.Q. 89.20 9.36 Correlation Coefficients WISC I.Q. TEST WAIS I.Q. Verb Perf F.Scale Verb Perf WISC Verbal 1.000 0.617 0.862 0.911 0.520 Performance 1.000 0.929 0.568 0.753 Full Scale 1.000 0.790 0.723 WAIS Verbal 1,000 0.597 Performance 1.000 Full Scale TEST F.Scale 0.419 0.669 0.750 0.735 0.805 1.000 TEST F.Scale 0.814 0.727 0.846 0.905 0.879 1.000 APPENDIX D SIGNIFICANCE OF THE DIFFERENCE BETWEEN I.Q. CORRELATIONS WITH GRADUATION AND READING INDEX CORRELATIONS WITH GRADUATION IN REGULAR CLASSES 167 168 EXHIBIT 1 SIGNIFICANCE OF THE DIFFERENCE BETWEEN I.Q. CORRELATIONS WITH GRADUATION AND READING INDEX CORRELATIONS WITH GRADUATION IN REGULAR CLASSES Formula ^*D = r f ’ z. - z ■ z 1 (Standard error of the difference be tween 2z coefficients) r— = I 2o8-3 .10 r = .602 r = .349 p < .001 r = .602 r = .031 p ^ .001 r = .602 r = .223 Reading Index - graduate status Verbal I.Q. - graduate status • (subtracting) remainder .32 = 3.2 (S.D.'s) .10 Reading Index - graduate status Performance I.Q. - graduate status (subtracting) ' remainder .66 = 6.6 (S.D.'s) 7m Reading Index - graduate status Full Scale I.Q.' - graduate status (subtracting) ' remainder 1 1 z = .69 z = .37 = .32 z = .69 z = .03 = .66 z = .69 z = .22 .47 .47 = 4.7 (S.D.'s) TTO p ^ .001 169) EXHIBIT 2 SIGNIFICANCE OF THE DIFFERENCE BETWEEN (1) I.Q. CORRELATIONS WITH GRADUATION IN EMR CLASSES AND (2) READING INDEX CORRELATION WITH GRADUATION IN EMR CLASSES Formula tfDz = <fZl - z2 N± -"5 N < ; I 1 159" 1 W .11 Reading Index - graduate status r = .328 z = .34 Verbal I.Q. - graduate status r = .194 z = .19 (subtracting) remainder . 15 .15 = 1.363 (S.D.'s) not significant .11 Reading Index - graduate status r - .328 z * .34 Performance I.Q. - graduate status r = .188 z = .19 (subtracting) remainder .15 .15 = 1.363 (S.D.'s) not significant - .11 Reading Index - graduate status r = .328 z = .34 Full Scale I.Q. - graduate status r = .229 z = .23 .11 = 1.0 (S.D.) not significant .11 .11 B I B L I O G R A P H Y 170 171 Ainsworth, S. H. "An Exploratory Study of Educational, Social and Emo tional Factors in the Education of Mentally Retarded Children in Georgia Public Schools," U. S. Office of Education, Cooperative Research Program, Project No. 171, Athens: University of Georgia Press, 1959. Alderdice, E. T., and Butler, A. J. "An Analysis of the Performance of Mental Defectives on the Revised Stanford-Binet and the Wechsler Bellevue," American Journal of Mental Deficiency, 56 (1952), 609-614. Alper, Arthur E. "An Analysis of the WISC with Institutionalized Re tardates," American Journal of Mental Deficiency, Vol. 71, No. 4 (January, 1967), 624-630. Alper, T. G., Boring, E. G. "Intelligence Tests of Northern and South ern White and Negro Recruits in 1918," Journal of Abnormal Psychology, 39 (1944), 457-471. Altus, Grace T. "A Note on the Validity of the Wechsler Intelligence Scale for Children," Journal of Consulting Psychology, 16 (1952), 231. Altus, Grace T. "WISC Patterns of a Selected Sample of Bilingual School Children," Journal of Genetic Psychology, 83 (1953), 241-248. Altus, Grace T. "A WISC Profile for Retarded Readers," Journal of Consulting Psychology, 20 (1956), 155-156. Anastasi, Anne. Some Implications of Cultural Factors for Test Con struction. Proceedings, 1949 Conference on Test Problems, Educational Testing Services, 1950, 13-17. Anastasi, Anne, et al. "A Comparison of Negro and White Preschool Children in Language Development and Goodenough Draw-a-Man I.Q." Journal of Genetic Psychology, 81 (1952), 147-165. Anastasi, Anne, and Cordova, F. A. "Some Effects of Bilingualism Upon the Intelligence Test Performance of Puerto Rican Children in New York City," Journal of Educational Psychology, 44 (1953, 1-19. Anastasi, Anne. Psychological Testing, New York: The Macmillan Com pany, 1954, 1959. Anastasi, Anne. Differential Psychology. New York: The Macmillan Company, 1958. Anastasi, Anne. "Cultural Differences," in C. W. Harris (ed.), Encyclo pedia of Educational Eesearch. New York: Macmillan, 1960. 172 Anderson, E. E., et al. "A Comparison of the Wechsler-Bellevue, Revised Stanford-Binet, and ACE Tests at the College Level," Journal of Psychology3 14 (1942), 317-326. Anderson, J. M. "Review of the WISC," in 0. K. Buros (ed.), Fourth Mental Measurement Yearbook3 Highland Park, N. J.: Greyphon, 1953. 480-481 . Angelino, H., and Shedd, C. L. "An Initial Report of a Validation Study of The Davis-Eells Test of General Intelligence on Problem- Solving Ability," Journal of Psychology, 40 (1955), Appell, M. J., Williams, F. M., and Fishell, K. N. "Significant Factors in Placing Mental Retardates from a Workshop Situation," Person nel and Guidance Journal3 41 (1962), 260-265. Arlitt, Ada H. "The Relation of Intelligence to Age in Negro Children," Journal of Applied Psychology 6 (1922), 378-384. Aimitage, S. G. and Pearl D. 'Wechsler-Bellevue Changes Over Time," . Journal of Clinical Psychology3 14 (1958), 22-24. Asher, E. J. "The Inadequacy of Current Intelligence Tests for Testing Kentucky Mountain Children," Journal of General Psychology, 46 (1935), 480-486. Atchison, C. 0. "Use of the WISC with 80 Mentally Defective Negro Children," American Journal of Mental Deficiency 3 60 (1955), 378-379. Babbott, E. F., and Grant, C. L. "I.Q. As One of Several Variables in Predicting Academic Success," The School Counselor3 120(1964), 18-21. Balinski, B. "An Analysis of Mental Factors of Various Age Groups from Nine to Sixty," Genetic Psychological Monographs3 23 (1949), 191-234. Bailer, Warren R. "A Study of the Present Social Status of a Group of Adults Who, When They Were in Elementary Schools, Were Classified as Mentally Deficient," Genetic Psychological Monographs3 18 (1936), 165-244. Banks, A. C. "A Comparative Study of Mentally Bright and Mentally Dull Negro High School Seniors," Dissertation Abstracts3 15, (1955). Barratt, E. S., and Baumgarten, Doris L. "The Relationship of the WISC and Stanford-Binet to School Achievement," Journal of Consulting Psychology3 21 (1957) 144. I 173 Baumeister, A. A. and Bartlett, C. J. "A Comparison of the Factor Structure of Normals and Retardates on the WISC,” Amevican Journal of Mental Deficiency, 66 (1962), 641-646 (a) . Baumeister, A. A. and Bartlett, C. J. "Further Factorial Investigations of Mental Defectives," American Journal of Mental Deficiency 3 67 (1962), 257-261 (b). Baumesiter, A. A., Bartlett, C. J., and Hawkins, W. F. "Stimulus Trace as a Predictor of Performance," American Journal of Mental De ficiency , 67 (1963), 726-729. Baumeister, Alfred A. , r Use of the WISC with Mental Retardates: A Review, American Journal of Mental Deficiency, 69 (1964965), 183-193. Baumeister, Alfred A. "Problems in Comparative Studies of Mental Re tardates and Normals," Amevican Journal of Mental Deficiency3 71 (March, 1967), 869-875. Bayley, Nancy. "On the Growth of Intelligence," Amevican Psychologist, 10 (1955), 805-818. Bayley, Nancy. "Data on the Growth of Intelligence Between 16 and 21 Years as Measured by the Wechsler-Belleivue Scale," Journal of Genetic Psychology3 90 (1957), 3-15. Bean, L. "Negro Responses to Verbal and Non-verbal Materials," Journal of Psychology3 13, 343-350. Bensberg, G. J. "The Relation of Academic Achievement of Mental De fectives to Mental Age, Sex, Institutionalization and Etiology," Amevican Journal of Mental Deficiency3 (1953), 327-330. Binet, A., and Simon T. The Development of Intelligence in Children. Translated by E. S. Kite. Baltimore: Williams § Wilkins, 1916. Birren, J. E., and Morrison, D. F. "Analysis of WAIS Subtests in Rela tion to Age and Education," Journal of Gerontology,i. 16 (1961), 363-368.' Blackman, L. S. and Heintz, P. "The Mentally Retarded," in Hunt, J. T. (ed.) Review of Educational Research, I, No. 1 (February,1966), 5. Blatt, B. "The Physical Personality and Academic Status of Children Who Are Mentally Retarded Attending Regular Classes," Amevican Journal of Mental Deficiency3 62 (1958), 810-818. 174 Bloom, B. S. Stability and Change in Human Characteristics. New York: John Wiley § Sons, Inc., 1964. Bobroff, Allen. "Economic Adjustment of 121 Adults Formerly Students in Classes for Mental Retardates," American Journal of Mental Deficiency3 60 (1956), 525-35. Boots, W. E. "A Study in the Intelligence of White and of Colored Grade School Children, High School Students, and College Fresh men." Unpublished Master's Thesis, University of Wisconsin, 1926. Boynton, PJ L. Intelligence3 Its Manifestation and Measurement. New York: Appleton, 1933. Brigham, C. A Study of American Intelligence. Princeton: Princeton University Press, 1923. Brown, F. "An Experimental and Critical Study of the Intelligence of Negro and White Kindergarten Children," Journal of Genetic Psychology3 65 (1944), 161-175. Bruce, M. "Factors Affecting Intelligence Test Performance of Whites and Negroes in the Rural Southi," Archives of Psychology3 No. 252 (1940). Buros, 0. K. The Third Mental Measurement Yearbook. New Brunswick, N. J.: Rutgers University Press, 1949. Buros, 0. K. (ed.) The Fifth Mental Measurements Yearbook. Highland Park: The Gryphon Press, 1959. Buros, 0. K. The Sixth Mental Measurements Yearbook, p. Highland Park: The Gryphon Press, 1965. Burt, Cyril. "The Evidence for the Concept of Intelligence," British Journal of Educational Psychology, 25 (1955). Burt, Cyril. "Intelligence and Social Nobility," British Journal of Statistical Psychology3 14 (1961), 3-24. Cain, L. F., and Levine, S. "The Mentally Retarded]1 in Michael, William B. (ed.) 1963 Review of Educational Research, Vol. 33, No. 1, 70-82. Campbell, D. J. and Stanley, J. C. "Experimental and Quasi-Experimental Designs for Research on Teaching," in N. L. Gage (ed.) Handbook of Research in Teaching. Chicago: Rand McNally, 1963, 171-246. 175 Canady, Heiman G. 'The Methodology and Interpretation of Negro-White Mental Testing," School and Society, Vol. 55, No. 1430 (May 23, 1942), 569-575. Canady, H. G. "The Problem of Equating the Environment of Negro-White Groups for Intelligence Testing in Comparative Studies," Journal of Social Psychology, 17 (1943), 3-5. Carson, A. S., and Rabin, A. I. 'Verbal Comprehension and Communica tion in Negro and White Children," Journal of Educational Psycho logy, (i960), 51. Cassidy, Viola M., and Phelps, Harold R. Postschool Adjustment of Slow Learning Children: A Study of Persons Previously Enrolled in Special Classes in Ohio. Bureau of Special and Adult Educa tion, Ohio State University Press, 1955. Cassidy, V. M. and Stanton, J. E. An Investigation of Factors Involved in the Educational Placement of Mentally Retarded Children: A Study of Differences Between Children in Special and Regular Classes in Ohio. U. S. Office of Education Cooperative Research Program Project No. 043, Columbus: Ohio State University, 1959. Cellura, A. Raymond, and Butterfield, Earl C. "Intelligence, the Ben- der-Gestalt Test, and Reading Achievement," American Journal of Mental Deficiency, 71 (1966), 60-63. Chapanis, A., and Williams, W. C. "Results of a Mental Survey with the Kuhlmann-Anderson Intelligence Tests in Williamson County, Tennessee," Journal of Genetic Psychology, 1945. Charles, Dan C. "Ability and Accomplishments of Persons Earlier Judged Mentally Deficient," Genetic Psychological Monographs, 47 (1953), 3-71 Charters, W. W. "Social Class and Intelligence Tests," in W. W. Charters, and N. L. Gage (eds.), Readings in the Social Psycho logy of Education. Boston: Allyn and Bacon, 1963, 12-21. Clarke, A. D. B. "The Measurement of Intelligence: Its Validity and Reliability" in Clarke, Anne M. and Clarke, A. D. B. (eds.) Mental Deficiency: The Changing Outlook. Glencoe: The Free Press, 1961. Clarke, D. P. "Stanford-Binet Scale 'L' Response Patterns in Matched Racial Groups," Journal of Negro Education, 10 (1941), 230-238. 176 Coers, W. C. "Comparative Achievement of White and Mexican Junior High School Pupils," Peabody Journal of Education; 12 (1935), 157-62. Collinann, R. D., and Newlyn, D. "Employment Success of Mentally Dull and Intellectually Normal Ex-pupils in England," American Journal of Mental deficiency; 61 (1957), 484-90. Conklin, R. C., and Dockrell, W. B. "The Predictive Validity and Sta bility of WISC Scores Over a Four Year Period," Psychology in the Schools; Vol. 4, No. 3 (1967), 263-266. Cooper, J. G. "Predicting School Achievement for Bilingual Pupils," Journal of Educational Psychology, 49 (1958) 31-36. Craig, Aiine L. "A Study of the Performance of Mexican Children on the Leiter International Performance Scale," unpublished Master's Thesis, Los Palgeles: University of Southern California, 1938. Cronbach, L. J., and Meehl, P. E. "Construct Validity in Psychological Tests," Psychological Bulletin; 52 (1955), 281-302. Cronbach, L. J. Essentials of Psychological Testing; 2nd ed. New York: Harper § Row, 1960. Daly, W. C., and Lee, R. H. "Reading Disabilities in a Group of Men tally Retarded Children: Incidence and Treatment," Training School Bulletin; 57 (1960), 85-93. Darwin, Charles. Origin of the Species. London: Murray, 1859. Davenport, E. Lee, "The Intelligence Quotients of Mexican and Non- Mexican Siblings," School and Society; 36 (September, 1932), 304-306. Davidson, K. S., et al. "A Preliminary Study of Negro and White Differ ences on Form 1 of the Wechsler-Bellevue Scale," Journal of Consulting Psychology; 14 (1950), 489-492. Davis, J. S. 'The Scatter Pattern of a Southern Negro Group on the Wechsler Intelligence Scale," Journal of Clinical Psychology; 13 (1957), 298-300. Davis, Leo J., Jr., "The Internal Consistency of the WISC with Mentally Retarded," American Journal of Mental deficiency3 Vol. 70, No. 5 (1966). Delp, Harold A. "Review of the WISC," in Buros, 0. K. (ed.) Fourth Mental Measurements Yearbook. Highland Park: Gryphon, 1953. 177 De Milan, J. "Bilingualism and the Wechsler Vocabulary Scale," Journal of Clinical Psychology, 21 (1965), 298. Deutsch, M. "The Disadvantaged Child and the Learning Process: Some Social, Psychological and Developmental Considerations," in Passow, A. H. (ed.) Education in Depressed Areas. New York: Columbia University, Teachers College Bureau of Publications, 1963, 163-179. Deutsch, Martin, and Brown, Bert. "Social Influences in Negro-White Intelligence Differences," Journal of Social Issues, Vol. 20, No. 2 (April, 1964). Dillon, H. J. Early School Leavers: A Major Educational Problem. New York: National Child Labor Committee, 1949. Dimichael, S. G. (ed.) Vocational Rehabilitation of the Mentally Retarded. Washington: U. S. Office of Vocational Reahbili- tation, 1950. Doll, E. A. The Measurement of Social Competence. Minneapolis: Educational Test Bureau, 1953. Dreger, Ralph M. and Riller, Kent S. "Comparative Studies of Negroes and Whites in the United States," Psychological Bulletin, 57 (1960), 361-402. Edwards, A. S., and Jones, L. "An Experimental and Field Study of North Georgia Mountaineers," Journal of Social Psychology, 9 (1938).' Edwards, A. L. Experimental Design in Psychological Research. New York: Rinehart, 1950. Eels, K., Davis, A., et al. Intelligence and Cultural Differences. Chicago: University of Chicago Press, 1951. Ellis, Norman R. (ed.) Handbook of Mental Deficiency. New York: McGraw-Hill, 1963. Estes, Betsy W. "Influence of Socioeconomic Status on Wechsler In telligence Scale for Children: An Exploratory Study," Journal of Consulting Psychology, 17 (1953), 58-62. Firik, S. L., and Shontz, F. C. "Inference of Intellectual Efficiency from the WAIS Vocabulary Subtest," Journal of Clinical Psychology, 14 (1958), 409-412. Fisher, G. M. "The Altitude Quotient as An Index of Intellectual Potential: I. WAIS Data for Familial and Undifferentiated 178 Mental Retardates," American Journal of Mental Deficiency 3 65 (1960), 252-255. (a) Fisher, G. M. "Differences in WAIS Verbal I.Q. and Performance I.Q. in Various Diagnostic Groups of Mental Retardates," Journal of Mental Deficiency3 65“ (1960), 256-260. (b) Fisher, G. M. "A Corrected Table for fttermining the Significance of the Difference between Verbal and Performance I.Q. 's on the WAIS and the Wechsler-Bellevue," Journal of Clinical Psychology 3 (1960). Fisher, G. M., Dooley, M. D., and Silverstein, A. B. , r WAIS Performance of Familial and Undifferentiated Mental Subnormals," Psychological Reports3 7 (1960), 268. Fisher, G. M. "Discrepancy in Verbal I.Q. and Performance I.Q. in Adolescent Sociopaths," Journal of Clinical Psychology3 17 (1961) 60. Fisher, G. M. "A Comparison of the Performance of Endogenous and Exogenous M R’s on the WAIS," Journal of Mental Deficiency Re search, 5 (1961), 111-114. (a) Fisher, G. M., Risley, T. R. and Silverstein, A. B. "Sex Differences in the Performance of Mental Retardates on the Wechsler Adult Intelligence Scale," Journal of Clinical Psychology3 17 (1961), 170. Fisher, G. M., Kilman, Beverly A., and Shotwell, A. M. "Comparability of Intelligence Quotients of Mental Defectives on the WAIS and the 1960 Revision of the Stanford-Binet," Jourtiallof Consulting Psychology3 1961, 192-195. Fisher, Gary M. "A Note on the Validity of the Wechsler Adult Intelli gence Scale for Mental Retardets," Journal of Consulting Psychology3 Vol. 26, No. 4 (1962). Fisher, G. M. "Further Evidence of the Invalidity of the Wechsler Adult Intelligence Scale for the Assessment of Mental Retar dates," Journal of Mental Deficiency Research3 6 (1962), 41-43. Fishman, Joshua A. (ed.) "Guidelines for Testing Minority Group Children," Journal of Social Issues3 Vol. 20 (1964). Fairbanks, Ruth F. "The Sub-Normal Child - Seventeen Years After," Mental Hygiene3 17 (1933), 177-208. Flescher, Irwin. "Needle Trade Opportunities for the Mentally Retarded," American Journal of Mental Deficiency3 61 (113-116) 1956. 179 Frandsen, A. N., and Higginson, J. B. MThe Stanford-Binet and the Wechsler Intelligence Scale for Children,” Journal of Consulting Psychology3 15 (1951), 236-238. Fry, Lois M. "A Predictive Measure of Work Success for High Grade Mental Defectives,” Amevican Journal of Mental Deficiency 3 '61 (1956), 402-408. Fulk, B. E. and Harrell, T. W. "Negro-White Army Test Scores and Last School Grade," Journal of Applied Psychology3 36 (1952). Garfield, S. L. "An Appraisal of Object Assembly on the Wechsler- Bellevue and WAIS," Journal of Clinical Psychology3 16 (1960) 8-9. Garretson, 0. K. "A Study of Causes of Retardation among Mexican Chil dren in a Small Public School System in Arizona," Journal of Educational Psychology, 19 (January, 1928), 31-40. Garrett, Henry E. "Negro-White Differences in Mental Ability in the United States," Scientific Monthly3 65 (1947), 329-333. Garrett, Henry E. Statistics in Psychology and Education. New York: Longmans, Green and Co., 1961. Garth, Thomas R. "Intelligence of Mexican School Children," School and Society3 27 (1928). Garth, Thomas R. and Johnson, Harper D. "The Intelligence and Achieve ment of Mexican Children in the United States," Journal of , Abnormal and Social Psychology3 29 (1934). Garth, Thomas R. et al. "Administration of Non-Language Tests to Mexicans" Journal of Abnormal and Social Psychology3 31 (April, 1936), 53-58. Gilliland, A. R. "Socio-economic Status and Race as Factors in Infant Intelligence Test Scores," Child Development, 22 (1951), 271-273. Gilmore, John V. Manual for the Gilmore Oral Reading Test. New York: Harcourt, Brace § World, Inc., 1951. Goddard, H. H. "What is Intelligence?" Journal of Social Psychology3 24 (1946), 51-69. Goldstein, Herbert. "Social Aspects of Mental Deficiency." Unpub lished Ph.D. Thesis. University of Illinois, 1957. ;180 Goldstein, Herbert, Moss, J. W. and Jordan, L. J. "The Efficacy of Special Class Training on the Development of Mentally Retarded Children," U. S. Dept, of Health, Education and Welfare, Office of Education, Cooperative Research Project No. 619. Urbana: Institute for Research in Exceptional Children, University of Illinois, 1965. Goodenough, Florence L. "Racial Differences in the Intelligence of School Children," Journal of Experimental Psychology, 9 (1926), 388-397. Graham, E. E., and Kamano, D. "Reading Failure as a Factor in the WAIS Subtest Patterns of Youthful Offdnders," Journal of Clinical Psychology, 14 (1958), 302-305. Green, R. L., et al. The Educational Status of Children in a District Without Public Schools. Cooperative Research Projects No. 2321. United States Office of Education, Department of Health, Educa tion and Welfare, 1964. Green, Robert L., et al. "Some Effects of Deprivation on Intelligence, Achievement, and Cognitive Growth," Journal of Negro Education, Vol. 36, No. 1, (Winter, 1967). Groelle, M. C. "Some Results and Implications of Reading Survey Tests Given to Educable Mentally Retarded Cyildren," Exceptional Children, 27 (1969), 443-448. Guertin, W. H., Frank, G. H., and Rabin, A. I. "Research with the Wechsler-Bellevue Intelligence Scale," Psychological Bulletin3 53 (1956), 235-257. Guertin, W. H. et al. "Research with the WAIS," Psychological Bulletin3 59 (1962), 1-26. Guertin, Wilson,H., Frank, George H., Ladd, Clayton E., Rabin, Albert I., and Hies ter, Douglas S. "Research with the Wechsler Intel ligence Scales for Adults," Psychological Bulletin, 66 (1966), 385-409, No. 5. Guilford, J. P. Fundamental Statistics in Psychology and Education. (3rd ed.) New York: McGraw-Hill, 1956. Guilford, J. P. "The Structure of the Intellect," Psychological Bulletin3 53 (1956), 267-293. Guthrie, G. M. and Pastovic, J. J. "Some Evidence of the Validity of the WISC," Journal of Consulting Psychology, 15 (1951), 385-386. 181 Haggard, E. A. et al. ' ’ Some Factors which Influence Performance of Children on Intelligence Tests," Amevican Psychologist, 3 (1948), 265-6 (Abstract). Hammer, E. F. "Comparison of the Performances of Negro Children and Adolescents on Two Tests of Intelligence, One as an Emergency Scale. Journal of Genetic Psychology, 84 (1954), 85-93. Hansen, Edith Davenport. "A Study of Intelligence Tests' Results for Mexican Children Based on English and Mexican Test Forms," Un published Master's Thesis, University of Southern California, 1931. Havighurst, R. J., and Breese, F. H. "Relations between Ability and Social Status in a Midwestern Community: III. Primary Mental Abilities," Journal of Educational Psychology, 38 (1947), 241-47. Havighurst, R. J., and Taba, Hilda. Adolescent Character and Person ality. New York: Wiley, 1949. Heber, Richard F. "A Manual on Terminology and Classification in Mental Retardation," American Journal of Mental Deficiency Monograph Supplements (L959} 64 No. 20. Hernandez, Luis F. "The Culturally Disadvantaged Mexican-American Student: Part I," Journal of Secondary Education, Vol. 42, No. 2 (February, 1967), 59-65. Hernandez, Luis F. "The Culturally Disadvantaged Mexican-American Student: Part II," Journal of Secondary Educations Vol. 42, No. 3 (March, 1966), 123-128. Hernandez, Luis F., and Johnson, Kenneth R. "Teaching Standard Oral English to Mexican-American and Negro Students for Better Voca tional Opportunities," Journal of Secondary Education, 42 (April 1967), 151-155. Hirsch, N. D. M. "An Experimental Study of the East Kentucky Mountain eers," Genetic Psychological Monographs, 3 (1928). Holland, W. R. "Language Barrier as an Educational Problem of Spanish- Speaking Children," Exceptional Childrens 27 (S. 1960), 42-50. Hopkins, K. "The Efficiency of the WISC in Predicting Organicity in Children with Normal I.Q. 's and a Note on Methodology." Un published study, 1960. Hopkins, Kenneth D. and Michael, William B. "The Diagnostic Use of WISC Subtest Patterns," California Journal of Educational Researchs XII, No. 3 (May, 1961), 116-117. 182 Horton, C. P., et al. "Growth and Development XI: Descriptive Analysis of the Backgrounds of 76 Negro Children Whose Scores Are Above or Below Average on the Merrill-Palmer Scale of Mental Tests at Three Years of Age, Journal of Genetic Psychology 3 100: 255-265. Hunt, J. McV. Intelligence and Experience. New York: The Ronald Press, 1961. Hutt, M. L. The Mentally Retarded Child. Boston: Allyn-Bacon, 1958. Jaffe, Jacob. "Attitudes of Adolescents Toward the Mentally Retarded," American Journal of Mental Deficiency3 70, No. 6 (May, 1966). Jastak, Joseph. Wide Range Achievement Test Manual. Wilmington, Del.: G. L. Story, Co., 1946. Jenkins, M. D. "A Socio-Psychological Study of Negro Children of Su perior Intelligence. Journal of Negro Education3 5 (1936), 175-190. Jenkins, M. D. "Case Studies of Negro Children of Binet I.Q. 160 and Above," Journal of Negro Education3 12 (1943). Jensen, Arthur R. "Learning Abilities in Mexican-American and Anglo- American Children," California Journal of Educational Research3 Vol. 12, No. 4 (September, 1961), 147-159. John, Vera P. "The Intellectual Development of Slum Children: Some Preliminary Findings Panel: Programs for the Socially Deprived," Urban Child3 1962. Johnson, G. 0. "Special Education for the Mentally Handicapped - A Paradox," Exceptional Children3 29 (October, 1962), 62-69. Johnson, G. 0. "A Study of the Social Position of Mentally Handicapped Children in the Regular Grades/' American Journal of Mental De ficiency , 55 (1950), 60-89. Johnson, G. 0. and Kirk, S. A. "Are Mentally Handicapped Children Segregated in the Regular Grades?" Journal of Exceptional Children, 17 (1950)', 65-68, 87-88. Jones, H. G. "The Evaluation of the Significance of Differences Be tween Scales Scores in the WAIS: The Perpetuation of a Fallacy," Journal of Consulting Psychology3 20 (1956), 319-320. Jurjevich, R. "An Analysis of the Altitude I.Q. 's of Delinquent Girls," Journal of General Psychology3 69 (1963), 221-226. Justman, Joseph. "Assessing the Intelligence of Disadvantaged Children," Education3 (February, 1967). Justman, Joseph. "Stability of Academic Aptitude and Achievement Test Scores of Mobile Disadvantaged Children." Unpublished paper. Kennedy, Ruby J. The Social Adjustment of Morons in a Connecticut City. Southbury Social Service Department, 1948. Kennedy, W. A., et al. The Standardization of the i960 Revision of the Stanford-Binet Intelligence Scale on Negro Elementary School Children in the Southeastern United States.' Cooperative Research Project No. 954, U. S. Office of Education, Department of Health, Education and Welfare, 1961. Kennedy, W. A. et al. "A Normative Sample of Intelligence and Achieve ment of Negro Elementary School Children inothe Southeastern United States," Monograph, Society for Research in Child Develop ment, 20, No. 6 (1963). Kephart, N. C. The Slow Learner in the Classroom. Columbus: Charles E. Merrill, 1960. Kern, William H., and Pfaeffle, Heinz. "A Comparison of Social Adjust ment of Mentally Retarded Children in Various Educational Settings," American Journal of Mental Deficiency 3 67 (November, 1962)', 407-13. Kirk, S. A. and Johnson, G. 0. Educating the Retarded Child. Boston: Houghton Mifflin, 1951. Kitzinger, Helen, et al. "Supplementary Guide for Administering and ’ Scoring the Wechsler-Bellevue Intelligence Scale Form I," Psychological Monograph3 65 (2) (1951), No. 319. Klineberg, 0. Race Differences. New York: Harper, 1935. Klineberg, 0. (ed) Characteristics of the American Negro. New York: Harper, 1944. Klineberg, 0., et al. "On Race and Intelligence: A Joint Statement," American Journal of Orthopsychiatry3 26 (1957), 121-134. Klineberg, 0. "Negro-White Differences in Intelligence Performance: A New Look at an Old Problem," American Psychologist3 18 (1963) 198-203. 184 Knapp, Robert T. "Effects of Time Limits on the Intelligence Test Performance of Mexican and American Subjects," Journal of Edu cational Psychology 3 51 (February, I960), 14-20. Knoblock, Hilda, et al. "Further Observations on the Behavioral Development of Negro Children," Journal of Genetic Psychology 3 83 (1953), 137-157. Koch, Helen L., et al. "A Study of the Test Performance of American, Mexican, and Negro Children," Psychological Monographs3 Vol. 35, No. 166, (1926). Kough, Jack, and DeHaan, Robert F. Teacher’ s Guidance Handbooks Vol. I: Identifying Children with Special Needs3 Chicago: Science . Research Associates, 1955. Krugman, J. I., Justman, J., Wrightson, J. W. and Krugman, M. "Pupil Functioning on the Stanford-Binet and the WISC," Journal of Consulting Psychology3 15 (1951), 475-483. Kureth, G., Muhr, J. P., and Weisgerber, C. A. "Some Data on the Validity of the WISC," Child Developments 23 (1952), 281-287. Lacy, L. D. "Relative Intelligence of White and Colored Children," Elementary School Journal3 (March, 1926), 542-546. Laing, Alice F., and Orazon, Maurice. "Sociometric Groupings Among Educationally Subnormal Children," American Journal of Menial Deficiency3 71 ‘ .(July, 1966), 73-77. Lambert, Nadine M. "The Present Status of the Culture Fair Testing Movement," Psychology in the Schools3 Vol. 1, No. 3, (July, 1964) 318-330. Lanfeld, E. S. and Saunders, D. R. "Anxiety as Effect of Uncertainty, An Experiment Illuminating the OA Subtest of the WAIS," Journal of Clinical Psychology3 17 (1961), 238-241. Lee, E. S. "Negro Intelligence and Selective Migration: A Philadel phia Test of the Klineberg Hypothesis," American Sociological Review3 16 (1951), 227-233. Lesser, Gerald S. "Mental Abilities of Children in Different Social and Cultural Groups3 Cooperative Research Project No. 1635. Coop erative Research Program of the Office of Education, U. S. Dept, of Health, Education and Welfare, 1964. Levinson, B. M. "A Research Note on Subcultural Differences in WAIS Between Italians and Jews," Journal of Gerontology3 15 (1960), 197-198. 185] Levinson, B. M. "Cultural Pressure and MIS Scatter in a Traditional Jewish Setting," Journal of Gene-bio Psychology, 93 (1958), 277-86. Levinson, B. M. "A Comparison of the Performance of Bilingual and Monolingual Native Bom Jewish Preschool Children of Traditional Parentage on Four Intelligence Tests, " Journal of Clinical Psycho logy, 15(1959), 74-76. ' Levinson, B. M. "Subcultural Values and I.Q. Stability," Journal of Genetic Psychology, 98 (1961), 69-82. Levinson, B. M. "Jewish Subculture and MIS Performance Among Jewish Aged," Journal of Genetic Psychology, 1002(1962), 55-68. Levinson, B. M. "The Wais Quotient of Subcultural Deviation," Journal of Genetic Psychology, 103 (1963), 123-131. Levinson, B. M. "A Comparative Study of the WAIS Performance of Native-Born Negro and White Homeless," Journal of Genetic Psychol ogy, (1964). Light, M. L., and Chambers, W. R. "A Comparison of the WAIS and Wechsler-Bellevue II with Mental Defectives," American Journal of Mental deficiency, 62 (1958), 878-881. Littell, William M. 'The Wechsler Intelligence Scale for Children: Review of a Decade of Research," Psychological Bulletin, Vol. 57 (March, 1960), 132-156. Lorge, I. "Schooling Makes a Difference," Teachers College Record, 46 (1945), 483-492. Maher, Brendan A. "Intelligence and Brain Damage," in Ellis, Norman R. (ed.) Handbook of Mental Deficiency, New York: McGraw-Hill, Inc., 1963, 224-252. Manuel, H. T., et al. "Intelligence and Drawing Ability of Young Mexican Children," Journal of Applied Psychology, 16 (August, 1932), 382-387. Maxwell, A. E. "Tables to Facilitate the Comparison of Subtest Scores on the WISC," Journal of Clinical Psychology, 15 (June, 1959), 293-295. McCandless, B. R. "Review of the WISC," in 0. K. Buros (ed.), Fourth Mental Measurements Yearbook. New Jersey: Gryphon Press, 1953. McCarthy, D. "Administration of Digit Span and Coding Subtests of the MIS and WISC to Left-Handed Subjects," Psychological Reports, 8 (1961), 407-408. 186 MeGurk, Frank. "On White and Negro Test Performance and Socioeconomic Factors," Journal of Abnormal and Social Psychology, Vol. 48, No. 3 (1953), 448-450. McNema, R. Q. The Revision of the Stanford-Binet Scale. Boston: Houghton Mifflin, 1942. McKeon, Rebecca M. 'Mentally Retarded Boys in War Time," Mental Hygiene, 30 (1946), 47-55. MacPhee, A. M. et al. "The Performance of Mentally Retarded Rural Southern Negroes on the Verbal Scale of the Wechsler Bellevue In telligence Examination," Journal of Social Psychology, 25 (1947), 217-229. Miklas, Milton J. "A Survey of Classes for the Educable Mentally Re tarded in Los Angeles County," California Journal of Educational Research3 XI, No. 4 (September, 1960), 176-178. Miller, D. R., Fisher, G. M., and Dingman, H. F. "A Note on Differen tial Utility of MIS Verbal arid PIQ’s. American Journal of Mental Deficiency, 65 (1961), 482-485. Miller, Robert V. "Social Status and SocioempatKic Differences Among Mentally Superior, Mentally Typical and Mentally Retarded Chil dren," Exceptional Children3 23 (1956), 114-119. Miner, J. B. Intelligence in the United States: A Survey with Con clusions for Manpower Utilization in Education and Employment. New York: Springer, 1956. Moore, J. E. "A Comparison of Negro and White Preschool Children on A Vocabulary Test and an Eye-Hand Coordination Test," Child Developments 13 (1942), 247-252. Murdy, W. G. "The Effect of Positive and Negative Administrations of Intelligence Test Performance." Dissertation Abstracts, 23 (1 9 62 ),' 1076 (Abstract). Mussen, P., Dean, S., and Rosenberg, Margery. "Some Further Evidence on the Validity of the WISC," Journal of Consulting Psychology3 16 (1952), 410-411. Newland, T. E., et al. "Chicago Non-Verbal Examination Results of an East Tennessee Negro Population," Journal of Clinical Psychology, 9 (1953), 44-46.' Newman, J. Robert, and Loos, Frank M. "Differences Between Verbal and Performance I.Q.'s with Mentally Defective Children on the WISC," Journal of Consulting Psychology} 19 (1955), 16. 187 Palomares, Uvaldo H., et al. "Evaluation of Mexican-American Pupils for EMR Classes," California Education3 3 (1966), 27-29. Pasamanick, B. and Knoblod, Hilda. "Early Language Behavior in Negro Children and the Testing of Intelligence," Journal of Abnormal and Social Psychology3 50 (1955), '401-402. Pasamanick, B. A. "The Contribution of Seme Organic Factors to School Retardation in Negro Children." Unpublished paper read before the APA in Chicago, September, 1956. Passow, A. H. (ed.) Education in Depressed Ares. New York: Teachers College Publications, 1963. Pastoric, J. J., and Guthrie, G. M. "Some Evidence on the Validity of the WISC," Journal of Consulting Psychology, 15 (1951), 385-86. Peal, Elizabeth, et al. "The Relation of Bilingualism to Intelligence," Psychological Monographs3 No. 27 (1962), 76. Peckham, Rolf A. "Problems in Job Adjustment of the Mentally Retarded," American Journal of Mental Deficiency, 56 (1951), 448-53. Peters, James S. "A Study of the Wechsler-Bellevue Verbal Scores of Negro and White Males," Journal of Negro Education3 29 (Winter, i960), 7-16. Pintner, R., and Paterson, R. G. Scale of Performance Tests. New York: Appleton § Co., 1925. Pintner, R., Intelligence Testing: Methods cmd Results. New York: Holt, 1931. Plant, Walter T., and Lynd, Celia. "A Validity Study and a College Freshman Norm Group for the WAIS," Personnel and Guidance Journal3 37 (April, 1959), 578-80. Plumb, Galen R. "Scoring Difficulty of Wechsler Comprehension Re sponses," Journal of Educational Psychology3 46 (1955), 179-183. Pressey, L. W. ’ The Influence of Inadequate Schooling and Poor En vironment upon Results with Tests of Intelligence. Journal of Applied Psychology3 4 (1920). Price, J. St. Clair. "Negro-White Differences in General Intelligence," Journal of Negro Education3 3 (1934), 424-452. Rapaport, D., et al. Diagnostic Psychological Testing. Chicago: Year Book Publishers, 1945, Vol. I., p. 573. 188 Reger, R. "Repeated Measurements with the WISC," Psychological Reports, 11 (1962), 418. Reger, Roger. "Myths About Intelligence," Psychology in the Schools,III (January, 1966), 39-43. Reger, Roger. "WISC, WRAT, and CMAS Scores in Retarded Children," American Journal of Mental Deficiency. 70 (March, 1966), No. 5 717-721. Robeck, Mildred C. "Subtest Patterning of Problem Readers on the WISC," California Journal of Educational Research, 11 (May, 1960) 110-115. Roen, S. "Personality and Negro-White Intelligence," Journal of Ab normal and Social Psychology, 61 (1960), 148-150). Sabagh, G., Dingman, H. F., Tarjan, G., and Wright, S. W. "Social Class arid Ethnic Status of Patients Admitted to a State Hospital for the Retarded," Pacific Sociological Review, 2 (1959), 76-80. Sandercock, Marjan G., and Butler, Alfred J. "An Analysis of the Per formance of Mental Defectives on the Wechsler Intelligence Scale for Children," American Journal of Mental Deficiency, 57 (1952), 100-105. Sarason, Seymour B. Psychological Problems in Mental Deficiency. (2nd ed.) New York: Harper, 1953. Sarason, S., and Gladwin, T. "Psychological and Cultural Problems in Mental Subnormality: A Review of the Research," Genetic Psycho logical Monographs, 57 (1958), 7-269. Satler, George C. "Psychometric Scatter Among Mentally Retarded and Normal Children," Training School Bulletin, 52 (1955), 63-68. Schwegler, R. A. "A Comparative Study of the Intelligence of White and Colored Children," Journal of Educational Research, (1920) 838-848. Seashore, H. A., et al. "The Standardization of the WISC," Journal of Consulting Psychology, 14 (1950), 99-110. Semler, I. J., and Iscoe, I. "Comparative and Developmental Study of the Learning Abilities of Negro and White Children Under Four Conditions," Journal of Educational Psychology, 54 (1963), 38-44. Shaffer, L. F. "Review of the WISC," Journal of Consulting Psychology, 13, (1949), 453-454. 189 Sheldon, Paul M. ' 'Mexican-Americans in Urban Public Schools: An Ex ploration of the Dropout Problem," California Journal of Educational Research, Vol. 12, No. 1 (January, 1961), 21-26. Sheldon, William H., "The Intelligence of Mexican School Children," School and Society, 19 (1924), 139-142. Sheperd, George. "What Research on the Mentally Retarded Has to Say," Journal of Secondary Education, 41 (December, 1966), 339-342. Sherman, M., and Key, Cora B. "The Intelligence of Isolated Mountain Children," Child Development, 3 (1932). Sherman, M. Intelligence and Its Deviations, New York: Ronald Press, 1945. • Shotwell, Anna M. "Arthur Performance Ratings of Mexican and Ameri can High-Grade Mental Defectives," American Journal of Mental Deficiency, 49 (1945), 445-449. Shuey, Andrez M. The Testing of Negro Intelligence. Lynchburg, Va.: Bell, 1958. Silverstein, A. B., et al. "Cultural Factors in the Intellectual Func tioning of the Mentally Retarded," American Journal of Mental Deficiency, Vol. 67, No. 3 (November, 1962), 396-401. Simon, L. M., and Levitt, E. A. 'The Relation between the Wechsler- Bellevue I.Q. Scores and Occupational Area," Occupations, 29 (1950), 23-25. Sparks, H. L. and Blackman, L. S. "What is Special About Special Education Revisited: The Mentally Retarded," Exceptional Children, 31 (January, 1965), 242-4.' Spearman, C. "General Intelligence," American Journal of Psychology, 15 (1904), 201-293. Stanton, Jeanette E. and Cassidy, Viola M. An Investigation of Factors Involved in Educational Placement of Mentally Retarded Children: A Study of Differences Between Children in Special and Regular Classes in Ohio. U. S. Department of Health, Education and Welfare, Office of Education, Cooperative Research Project, Number 043, Columbus: Ohio State University, 1959. Stewart, Naomi. "AGCT Scores of Army Personnel Grouped by Occupation, Occupations, 26 (1947), 5-41. Tanser, H. A. The Settlement of Negroes in Kent County, Ontario. Chatham, Ontario: 1939. 190 Tarjan, G., Dingman, H. F., Eyman, R. K., and Brown, S. J. "Effective ness of Hospital Release Programs." American Journal of Mental Deficiency3 64, (1960), 609-617. Teahan, J. F., and Drews, Elizabeth M. "A Comparison of Northern and Southern Negro Children on the WISC," Journal of Consulting Psychology3 26 (1962), 292. Terman, L. M. The Measurement of Intelligence. Boston: Houghton Mifflin, 1916. Terman, L. M. et al. Genetic Studies of Genius: The Mental and Physi cal Traits of a Thousand Gifted Children. Vol. I. Stanford: Stanford University Press, 1925. Terman, L. M., and Merrill, Maud A. Measuring Intelligence. Boston: Houghton Mifflin, 1937. Terman, L. M., and Merrill, Maud A. Stanford-Binet Intelligence Scale. Boston: Houghton Mifflin, 1960. Throne, F. M., Schulman, J. L., and Kayser, J. C. "Reliability and Stability of the WISC for a Group of Mentally Retarded Boys," American Journal of Mental Deficiency3 67 (1962), 455-457. Thurstone, Thelma S. An Evaluation of Educating Mentally Handicapped Children In Special and Regular Classes. U. S. Department of Health, Education and Welfare, Office of Education, Cooperative Research Program, Project Number 513. Grambling: Grambling College of Louisiana, 1960, 245 p. Tredgold, R. F., and Soddy, K. A Textbook of Mental Deficiency. Baltimore: Williams § Wilkins, 1956. Triggs, F. 0., and Carter, J. K. "Preschool Pupil Performances in the Stanford-Binet and the WISC," Journal of Clinical Psychology3 9 (1953), 27-29. Triggs, F. 0., Carter, J. K., Binks, Virginia, Foster, P. and Akanis, N. A. "The Relationship Between Specific Reading Skills and General Ability at the Elementary and Junior-Senior High School Levels," Educational and Psychological Measurement, 14 (1954), 176-185. Valett, Robert E. "Classification of the Mentally Retarded: An Appraisal," Psychology in the Schools3 Vol. 2, No. 3 (July, 1965), 210-213. 191 Webb, Allen P. MA Longitudinal Comparison of the WISC and WAIS with Editable Mentally Retarded Negroes,'! Journal ofi Clinical Psychology, 19 (1963), 101-102. Webb, Allen P. "Some Issues Relating to the Validity of the WAIS in Assessing Mental Retardation," Journal of Educational Research^ (May, 1964). Wechsler, D., et al. "A Study of the Subtests of the Wechsler-Bellevue Intelligence Scale in Borderline and Mentally Defective Cases," American Journal of Mental Deficiency3 45 (1941), 555-558. Wechsler, D. The Measurement of Adult Intelligence. (3rd ed.) Baltimore: Williams and Wilkins, 1944. Wechsler, D. Wechsler Intelligence Scale for Children. New York: Psychological Corporation, 1949. Wechsler, D. Manual for the WISC. New York: The Psychological Corporation, 1950. Wechsler, D. Manual for the Wechsler Adult Intelligence Scale. New York: The Psychological Corporation, 1955. Wechsler, D. The Measurement and Appraisal of Adult Intelligence. (4th ed.) Baltimore: Williams and Wilkins, 1958. Weider, A., Noller, P. A. and Schraum, T. A. "The Wechsler Intelligence Scale for Children and the Revised Stanford-Binet," Journal of Consulting Psychology3 15 (1951), 330-333. Wheeler, L. R. "The Intelligence of East Tennessee Mountain Children," Journal of Educational Psychology3 23 (1932) . Wiener, G. "The Effect of Distrust on Some Aspects of Intelligence Test Behavior," Journal of Consulting Psychology, 21 (1957) ,127-30. Yerkes, R. M. (ed.) Psychological Examining in the United States Army. National Academy of Science, 1921. Young, Florence M. et al. "Results of Testing Negro Rural Juveniles with the WISC," Journal of Social Psychology3 39 (1954), 219-26. Zach, Lillian. "Current Thought on Intelligence Tests," Psychology in the Schools, Vol. Ill,’ No. 2 (April,- 1966), 116-123.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Structure-Of-Intellect Factor Abilities Relating To Performance In Tenth-Grade Modern Geometry

PDF

A Descriptive Study Of Mature And Younger Women In An Associate Degree Nursing Program

PDF

Prediction Of Therapeutic And Intellectual Potential In Mentally Retardedchildren

PDF

Ego Diffusion In Women With Behavioral Disorders And The Integrating Effects Of Psychodrama In Identity Consolidation

PDF

A Comparative Analysis Of Three Approaches To Beginning Reading

PDF

Advance Organizers And The Enchancement Of Meaningful Verbal Learning Andretention

PDF

Evaluation Of Short-Term Training For Rehabilitation Counselors: Effectiveness Of An Institute On Epilepsy

PDF

The Relationship Of Self-Actualization Levels Of Counseled And Non-Counseled Parents To Perceptions Of Behavioral Characteristics In Their Children Previously Diagnosed As Disordered

PDF

A Comparison Of Two Instructional Programs For Severely Retarded Readers At The Junior High School Level

PDF

A Diagnosis Of The Management Styles Of High School Principals

PDF

A Factor Analytic Study Of Tests Designed To Measure Reading Ability

PDF

Disponibilite In The Therapeutic Relationship

PDF

Sound Blending: The Relationship Of Inter-Letter Interval, Consonant Category, And Stimulus Word Frequency To Word Identification

PDF

The Effects Of Einstellung On Visual Perception Of Random Shapes: A Multidimensional Scaling Analysis

PDF

Maternal Child-Rearing Attitudes And Developmental Growth Of Rubella Deaf-Blind Children

PDF

Value Orientations Of Parents Of Academically Successful And Unsuccessfulchildren

PDF

A Study Of The Factorial Validity And Reliability Of The Individual Test Of Creativity

PDF

Effects Of Biofeedback Training Of Alpha Rhythm On Arousal And Therapeutic Gain As An Adjunct To The Rehabilitation Of A Substance Dependent Population

PDF

The effect of noise on intellectual performance as related to personality and social factors in upper division high school students

PDF

Prescriptive Teaching As A Supplement To Behavior Modification In The Remediation Of Learning Disorders

Asset Metadata

Creator Simpson, Robert Lee (author)

Core Title A Study Of The Comparability Of The Wisc And The Wais And The Factors Contributing To Their Differences

Degree Doctor of Philosophy

Degree Program Education

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag education, educational psychology,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Schrader, Don R. (committee chair), Lovell, Constance (committee member), Meyers, Charles Edward (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-631587

Unique identifier UC11361273

Identifier 6817043.pdf (filename),usctheses-c18-631587 (legacy record id)

Legacy Identifier 6817043.pdf

Dmrecord 631587

Document Type Dissertation

Rights Simpson, Robert Lee

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA