Standardization of tests for defective children

Front matter

Itandardization of Tests for Defective Children

A DISSERTATION

■M. '1 ! I ! I i 1 ■ .hi . .

Table Of Contents

PAGE

I. Introduction ^

II. Historical development of tests of mental measur-

Introduction

The following study was undertaken to provide data for the standardization of the Healy-Fernald tests, described by the authors in Psychological Monographs, Vol. XIII, No. 2. It is offered because a standarization of these tests has been asked for by various persons interested in the development of clinical psychology. It is hoped that the study will also suggest some further clinical uses of the Binet-Simon series. The purpose of standardization is to show the reactions of a socially homogeneous group of individuals considered socially and pedogogically normal, classified according to certain principles discussed below.

The psychological considerations which underlie the author's classified evaluation of these tests has been found of value in clinics where the mentally defective must be distinguished from the mentally normal, and classified for the purposes of instruction in the public school and for placing in public institutions. It is hoped that the formulation of the principles which have led to such classifications may prove to be suggestive to those seeking further light on the process of mental diagnosis.

The study is offered, however, with a due appreciation of its incompleteness. It is desirable that a hundred or more children of each age and grade of the school should be given the tests discussed in the following pages, rather than the twenty which it was possible to get. But, though the numbers are small, they are sufficient to show clearly marked tendencies to specific types of reaction for the various grades and ages tested.

This work was done by the author while psychologist to the Juvenile Psychopathic Institute. This organization maintains a clinic at the Juvenile Court of Chicago. Wherever in this work "the clinic" is mentioned it is this clinic to which reference is made.

Much that is said in this study concerning mental tests in general and concerning many of the individual tests is the result of a body of experience gained in the clinic just mentioned and a further year's experience in the clinic of the Department of Child Study of the Board of Education of Chicago. Much of this experience is not amenable to statistical classification. The particular bit of work which is here used for standardization of the Healy-Femald tests is especially suitable for this purpose because it is gained from a group of socially homogeneous subjects.

The work of the Juvenile Court clinic is entirely with the juvenile delinquent; and except for this one characteristic that group of cases is not homogeneous. The work in the Department of Child Study is with children who for some reason are reacting unsatisfactorily to the school situation. This group of subjects is far from homogeneous mentally, physically and socially.

Since one of the reasons for atypical social response may be mental defectiveness, the reaction of what has come to be accepted by the users of these tests as mentally defective response has been compared with the mentally normal for each test. The factors which enter into consideration in classifying as mentally defective are discussed on page 164.

Subjects. — The children who served as subjects for the tests comprised the kindergarten and first six grades of a private school in Chicago. These children composed as perfectly homogeneous a group as it is probably possible to find in a school. They were the children of people of the professional class mainly. A few were children of successful business men who sought the best obtainable type of education for their children. The school was founded for the purpose of putting into application the broadest and best conceptions of educational theory and practice.

So far as heredity in its relation to social class is concerned these children were equally endowed. Home environment with reference to educational endowment and stimulus was uniform as nearly as such a matter may be measured. One may assume that the children who belonged to the same grade had had the some educational regime in home and school.

II Historical Development Of Tests Of Mental

MEASUREMENT IN CORRELATION WITH

GENERAL INTELLIGENCE

Psychologists have for some twenty years been concerned with finding a measure of general intelligence. Work with different types of tests under more or less rigorously controlled laboratory conditions has been carried on with children of different classes and with college students, graded in one way or another according to degrees of general intelligence. Within the last ten years there has arisen the need of application of the work of the laboratory psychologist to the practical work of clinics for investigation of socially atypical individuals. In the United States and other countries the criminal, the mentally defective, the backward school child and the supernormal child are being investigated with greater thoroughness than ever before. To meet this need there have been devised and invented tests for measuring mental ability of a type quite different from the tests devised by the early laboratory psychologists who worked at the problem. A short historical survey will serve to characterize the two groups of practical tests discussed in this article.

Only a part of the great mass of literature on tests of general intelligence is outlined in the following summary, namely, that part which endeavors to establish a measure for the general intelligence of children.

In 1897 the American Psychological Association (i) received the report of a committee appointed at its previous meeting to investigate the subject of physical and mental tests. This committee agreed upon and reported a series of tests which it recommended be tried on college students in the various psychological laboratories of the country. The physical and mental tests recommended were classified as follows :

Preliminary Data. — Date of birth; birthplace; birthplace, of father, birthplace of mother; occupation, including class in college; occupation of father; any measure previously made. Color of eyes; color of hair; right or left handed. Mother's maiden name; number of brothers; sisters; order of birth; age of parents at birth; birthplace and occupation of grandparents. Assymetry of body; color of eyes, hair, complexion; degeneracy or other stigma of head, eyes, ears, mouth, teeth, hands, feet, posture; gait; manners; coordination and speech; indications of intellectual, emotional and moral characteristics.

Physical Measurements. — Height, weight, and size of head. Breathing capacity. Height sitting.

Keenness of lAsion

Color vision

Keenness of hearing

Perception of pitch

Fineness of touch

Sensitiveness to pain

Perception of weight or force of movement

Dynamometer pressure of right and left hand

Rate of movement

Fatigue

Will power

Voluntary attention

Right and left hand movement

Rapidity of movement. — Taps on telegraph, short marks, trilling with two fingers or five.

Accuracy of aim

Reaction time for sound

Reaction time with choice. — Card sorting

Rate and discrimination of movement. — Marking out lOO a's in 500 letters, one of a number of geometrical figures, or colors, or pictures, or objects.

Quickness of distinction and movement. — Rate at which cards are sorted, combine with reaction, with choice, with effects of practice.

Perception of size. — Draw a line equal to a model 5 cm. in length, bisect it, erect a perpendicular of the same length, and bisect the right hand angle.

Perception of time. — The accuracy with which a standard time can be reproduced.

Memory. — The accuracy with which eight numerals heard once can be reproduced, and the accuracy with which a line drawn by the observer at the beginning of the hour can be reproduced at the end of the hour; Hne to be identified, not drawn; ten numerals ; nine numerals. A combined test of memory, association and finding time as described in the catalogue of the Columbian Exposition, accuracy of observation and recollection as proposed by Cattell and Bolton.

Memory type. — Variation in the use of ten numerals, compare results for indication of memory type and kind of imagery preferred.

Apperception test of Ebbinghaus

Imagery

Much, and more than was here recommended, of this work was done in the next few years, and the attempt was made to correlate the results with other evidences of general intelligence.

Before this time, however, some very important pieces of work had been done. One of these was the work of Gilbert (2) in 1894. Gilbert correlated weight, height, lung capacity, simple reaction time, reaction time with discrimination and choice, and time memory with mental ability. In determining general mental ability the teacher's judgment was relied upon. Each teacher classified her pupils as bright, average and dull. Approximately 100 children of each age group from six to seventeen were measured.

Porter (3), in 1893, correlated the height and weight of 33,500 school children with age and grade.

Later Smedley (4), in 1900, correlated height, standing and sitting; weight; ergograph and dynamometer records, and lung capacity with age and school standing of children between the ages of eight and eighteen inclusive.

West (5) correlated physical development and intellectual ability of Toronto school children. His method of grouping for intellectual ability was according to the teacher's estimate of the children as good, average, and poor.

The investigations of the men mentioned and others gave rise to three different conclusions with regard to the relation of physical development and mental ability. Porter, Smedley and some others found a positive correlation between physical development and mental ability of which success in school life was taken as the measure. Gilbert found no correlation and West found a negative correlation. These differences were probably due to different arrangement of data and to different methods of grouping grades of intellectual ability. Gilbert did not separate the sexes in his tables and did not state the proportions of his dull and bright groups. Gilbert and West classified according to the teacher's judgment; Porter and Smedley according to school grade with reference to age. A severe criticism of the method of classification according to teacher's judgment may be quoted from West's account of his experience with it :

"It soon became apparent to me that any such classification of children's mental ability would be very greatly influenced by the mental caliber of the teacher making the classification. . . . There were no poor scholars. The teachers were perfectly willing to classify the scholars as of good and average intelligence but any intimation of poor or stupid scholars was taken as a personal reflection upon the teacher of the class in question. . . . The poor students were no more than a mere handful."

The method of grouping according to success in school life, the method of Porter and Smedley, grades all children of the same age according to the same standard. Though some allowance must be made for a small group compelled by individual circumstances to residence in grades below that of which the mental ability of each individual might otherwise permit.

In the reaction time tests Gilbert thought to have found a correlation with intelligence. He says, "The curves for reaction time gave the most positive results showing that the brighter the child the more quickly he is able to act. In discrimination the same relation is true but to a less degree. ... Of time memory it may be said in general that the brighter the child the more accurate his sense of time." An examination of Gilbert's tables, however, fails to support so optimistic a view of the existing correlation. At some ages the dull class is superior to the bright class, and the differences between the three classes measured are everywhere slight.

Reaction time tests of various kinds were tried out by various investigators in the following years. The results of this work were reviewed by Whipple (6) in 1904. He distinguished between two kinds of reaction time experiment, the laboratory type, and the anthropometric type which includes card sorting and similar tests. The former consists of work done under rigid laboratory conditions upon subjects competent to make introspective analyses. The latter consists of experiments made on children or others without introspective analyses and without practice in direaction of attention. He pointed out that because different experimenters obtained widely varying results under the same conditions, attention came to be directed to the different types of individuals, and the reaction time experiment came to be an experiment in reaction consciousness. The anthropometric type of experiment he criticised, because the conditions under which the data were supplied, were so loosely controlled that one could not be sure what was measured. He concludes that any reaction time is conditioned upon a large number of independent factors and when these are eliminated or controlled in the laboratory "we have left no residuum of individual variation that can be turned to account in estimating the observer's general intelligence or mental ability."

In 1901 Wissler (7) pubHshed the results of a long series of tests and anthropometrical measurements made under the direction of Cattell upon students of Columbia University for a period of seven years. The results of these tests were correlated with class standing. The general conclusions were that the laboratory mental tests show little correlation in the case of college students ; that the physical tests show a general tendency to correlate with themselves but only to a very slight degree with mental tests; that the markings of students in college classes correlate with themselves to a considerable degree but not with the tests made in the laboratory.

Griffing (8), 1895, and others investigated the subject of attention with reference to general intelligence. In general some form of the tachistoscope was used. Griffing's conclusions represent the general concensus of opinion among these investigators. He said, "I found that those rated 'A' for mental capacity by the teachers on an A B C basis, had somewhat higher averages than the others. . . . There are, however, marked exceptions. . . . Those marked 'A' by their teachers for attention in class also excelled the others, but here also I found decided exceptions. Manypupils must have, therefore, good powers of attention, even when they show no evidence of them to their teachers."

In 1904 Spearman (9) made a critical analysis of the methods of work in the determination of correlations of various tests with general intelligence. With a more exact mathematical formula for the calculation of correlations and by the use of more factors for the determination of general intelligence he found large correlations in tests of discrimination of grays, and weight and pitch, with general intelligence.

In 1909 Burt (10) correlated the general intelligence of two sets of English schoolboys with tests of discrimination of two points upon the skin, of lifted weights, of pitch, and of length of lines. To these he added two motor tests, tapping and card dealing; two sensori-motor tests, card sorting and alphabet finding; tests of immediate memory of concrete words, abstract words and nonsense syllables ; the tracing of a geometrical pattern seen in a mirror, a test of the power to acquire new coordinations; the reproduction from memory of a pattern of spots presented by the tachistoscope upon squared paper; and a test of voluntary attention, which consisted of pricking an irregular line of dots passing rapidly before the subject.

Great care was exercised by Burt, in accordance with the recommendations of Spearman, in the mathematical work of correlating the test findings with general intelligence. The latter was estimated by the headmasters of the schools from which the reagents came. The conclusions at which Burt arrived are as follows :

Of the simple sensory tests, tactile, weight, pitch, and length of line discrimination, he says, "There appears to be no general connection between intelligence and capacity to discriminate weights ; any connections between intelligence and tactile discrimination, if it exists, is of the slightest; there is considerable connection between intelligence and capacity to discriminate undoubted general connection between intelligence and visual discrimination of lengths. ..."

Of the motor tests, tapping and card dealing he says : "Motor tests seem to have a higher correlation with intelligence than sensory tests. But where rapidity is due to frequent practice. . . the correlations with intelligence and other tests are reduced, abolished, or inverted. Thus so far as motor rapidity is the function of temporary 'facilitation' of the paths of neural discharge it appears also to be a function of intelligence, while so far as it is a function of permanent 'canalisation' of those paths it is but slightly or inversely related to intelligence." This latter conclusion was the result of disturbance of correlational results among a group of practiced card players.

Of the two sensori-motor tests, one for the sorting of cards according to color, the other for selecting a complete alphabet from a mixture of two alphabets, Burt says : "Depending as they do for their perfomiance upon processes of a more complex nature and a higher mental level, tests combining perception with motor reaction seem to involve the intelligence to a still higher degree than relatively simple sensory or motor tests. Of the two above discussed the alphabet seems to be, in practice, far the more efficient."

Immediate memory was tested by the use of sets of words of abstract significance, of concrete significance; and nonsense syllables. The correlation with general intelligence found by Burt between concrete memory, abstract memory, and nonsense memory was .58, .48, .43 respectively for one group of boys and .84, .78, .75 for another group. "Thus the memory for abstract words does not show a higher, but a lower correlation, with intelligence. . . . The introduction of difficult vocables, whether abstract nouns or meaningless syllables, proves in both groups to be on the whole a distracting element."

In the mirror test, a pattern is traced which, with the hand doing the work, is seen only in a mirror. This tests the subject's ability to readjust certain already learned eye and hand coordinations to a changed situation. Burt found many difficulties in the mechanical operation of this test as well as in the method of measuring results. In his judgment it was a test which with further perfection would be of great practical value. He, howlo CLARA SCHMITT

ever, sums up the factors which would compHcate and make uncertain the resuhs of this test in practice when he accounts for the divergence between his two groups of the correlational figures, .67 and .54. "The divergence between the two schools is largely due to the fact that four of the Preparatory schoolboys had had previous practice at an analogous task in the form of a not very common parlor pastime. Only one of the thirty Elementary boys had done any similar exercise before. The divergence might also be in part attributed to a greater familiarity with the use of the mirror among boys of a higher class as compared with boys of a lower status. A similar factor apparently operated when the test was applied to children of the opposite sex, though subsequent application to very young children, and to adults, have led me to wonder whether we are not dealing with one of the uninvestigated innate differences between the sexes."

The spot pattern test was given in a dark room by means of a tachistoscope. The pattern was shown as many times as it was necessary for the subject to learn to reproduce it correctly. The difficulties in the use of the tachistoscope in practical work are indicated when he says, "The tachistoscope was found to require a larger amount of experience on the part of both subjects and operators, than any of the other tests, except perhaps those involving sensory discrimination. . . . The first series of all had to be rejected as worthless, owing partly to the irrelevant excitement aroused in the subjects by the 'electrical flash' as the boys named it. . . . At this school we were not able to obtain the complete darkness and silence procured at the other in our extemporised dark room, and consequently the reliability coefficient and the raw correlations with intelligence are not so high." The coefficient obtained was .y6 and .75 for the two groups.

Burt's test of sustained attention consisted of pricking dots irregularly arranged upon a strip of paper which passed before the subject. The number of dots per minute which the student marked constituted a measure of his ability. The correlation for this test with general intelligence was found to be .75 and .96 for the two groups of boys.

STANDARDIZATION OF TESTS ii

Of the practical significance of these tests Burt says : "Of the twelve tests six furnish coefficients below .50 and six above .50. The former six — the simple sensory and motor tests — are thus of little use in the empirical diagnosis of intelligence. Among the latter six, no single test, at any rate in its present form, can be claimed as a self-sufficient instrument for measuring and detecting ability in individuals. But they indicate the direction in which such a test may hopefully be sought. . . . McDougall's dotting machine seems to be the most scientific. Where the external conditions could be kept most uniform, . . . both the amalgamated and the average raw coefficients reached .84. Such uniformity is difficult in more extensive work, and the ensuing variety in attention and fatigue affect the performances with this test. Moreover, its figures are less discriminative than either of the other three. By increasing the number of spots in the pattern the tachistoscope test may be made to differentiate with almost any degree of minuteness. . . . It is a slow test, however, and without repetition scarcely reliable. And it calls for some experience both on the part of the boys to grasp the nature of the task, and on the part of the experimenter to manipulate the apparatus with regularity. . . . The mirror test can be procured with but little trouble and expense, and needs no trained superintendent. It, too, requires further improvements, especially in procedure and calculations, to eliminate the influence of possible previous practice, and to elicit more completely the significance of the figures observed. If called upon to recommend a simple test for immediate use upon untrained subjects, I should be inclined to advocate the alphabet test as perhaps the simplest and most satisfactory test of all."

The work of Burt has been so fully recorded because it is the broadest and most careful attempt to correlate the results of tests with general intelligence. In this work many considerations were taken into account in constructing a scale of general intelligence of the subjects. Great care was taken in the management of the tests themselves aridi the mathematical correlations were worked out with accuracy. The conclusions drawn of the various tests are valuable from the standpoint of an interpretation of intelligence. For various reasons, however, the results can not be immediately applied to clinical work. The reasons maybe grouped under five heads.

The first and most important is the measurement of the tests against time. It is the experience of the writer and others in the work of children's clinics that time within the limits of rigid laboratory procedure can not be taken as the measure of the subject's ability with a particular test. This is because of the peculiar demands of such a clinic. One wishes the child to be unaffected by any feelings of fear or anxiety or strangeness with the situation when he comes up for examination. As far as possible he should not know that he is being examined. He, therefore, should not be subjected to the anxious desire to make good time in any thing he is doing. There are few tests which can at all lend themselves to such measurement, since in any test which can be useful as a measure of intelligence or which can show the child's intelligence functioning, there are involved perceptual or other types of discrimination which may be interfered with if the child anxiously desires to make a good time record. This interference with thought processes may cause the final result to be a misleading and perhaps unfair judgment of his general intelligence. Only such discriminations as are habitual with him and therefore make little demand upon attention can be measured against time. Under such circumstances one does not know what is being measured.

The second reason for the impracticabihty of Burt's work is the fact that with tests which correlate most highly with general intelligence, the use of apparatus is necessary. Burt, himself, showed how the tachistoscope mechanism interfered with attention to the object of the test itself. In much larger measure would such a piece of apparatus be a stumbling block in a clinic where defective individuals are examined. As was remarked above, one wishes the child to be unaffected by any feelings of fear or strangeness with the new and usually strange situation into which he is thrust when he comes to a psychological clinic for examination. A piece of strange apparatus will so

STANDARDIZATION OF TESTS I3

fill his mind with fear or speculation as to its object and its affect upon himself, that he may not act normally. The fact also that much practice is needed on the part of the subject to use a piece of apparatus must put it out of consideration for the practical clinician.

The third reason for discarding some of these tests is that previous practice may vitiate results. What Burt has said with regard to the mirror test and card dealing test has already been quoted. No tests except those on reading, arithmetic or other subjects of cultural value can be used in a practical clinic which in any way may have been the subject of practice on the part of the child, and these only if the examiner knows the extent of the child's experience with them.

The fourth point of error in applying the work of Burt, and all other works of its type, to the examination of atypical children is that it does not take into account the child's motive for action. In child life there are in general two motives for voluntary activity. One of these is the play motive, in which the child voluntarily seeks the end to be attained. The other is the social motive of pleasing associates, who may demand ends which he would not voluntarily seek to attain. In a psyschological laboratory the motives of the adult subject and the observer are the same, — the production of scientific data. In such case the subject lends himself willingly to any conditions imposed. Knowing the end of the experiment he is able to direct attention to the attainment of that end and away from the distracting elements of unusual conditions. It is otherwise with the child. With him the motive most conducive to natural reaction, uncomplicated by disturbing emotions, is the play motive.

The fifth reason for the lack of apphcability of this type of work to a practical clinic is that it has been done with subjects of so high a type of mentality that the results are of little value in measuring low types. A mathematical statement of the correlations of a test to the general ability of such subjects as form the reagents for experiments under laboratory conditions can have little significance in a clinic for defective or abnormal individuals. The subjects for laboratory experiments are in general of a high type of intelligence. Among Burt's subjects there was only one defective child. In a clinic no child typical of the average in the social situation in which he finds himself, ever comes up for examination. Necessarily in some realm of social, mental or physical functioning, the child to be examined is abnormal or he would not be brought for examination.

A part of Burt's work, however, can be of greater value for clinic purposes than his correlations led him to believe. Such simple sensori-motor tests as card sorting, or as Whipple designated them, the anthropometrical reaction tests, have been found by him and other laboratory experimenters to correlate little with general intelligence. The reason for this is that even with the least intelligent subjects these tests fall well within the limit of their intelligence, and, therefore, can not form a measure of the mental ability of those particular subjects. With many subjects of the practical clinic, however, such simple tests may be of great value. If it is found that the child can do nothing more complex than sort cards according to color or geometrical form, or whatever type of discrimination is employed, this, in part at least, establishes a measure of his mental ability. If it is found that he can not do even these tests then others of a simpler nature still must be used as a measure.

Viewed from the standpoint of such use, the simple "anthropometrical reaction" tests are of great value as measures of certain elements of general intelligence. For use in a practical clinic any test may serve as a measure, in whole or in part, if it really does mark off a range of intellectual activity. However, no test can be of such general use as Burt thought possible of the alphabet sorting test which he considered most valuable as a test of mental measurement. The highly intelligent child can accomplish tests of far greater complexity; the low grade child may be unable to make such fine discriminations. This test, therefore can not serve as a measure of these two grades of mental ability.

STANDARDIZATION OF TESTS I5

To sum up, the clinic must discard time for the most part as an important factor in the measurement of results. This point is discussed further below in connection with certain tests. Some other measure, preferably a qualitative one, must be substituted for this quantitative one. The clinic must discard rigorous laboratory conditions and adjust its tests to conditions more in conformity with those of everyday life. It must discard such apparatus as requires practice on the part of the siibject, or as is not directly connected with the object of the test.