I'M -ii ;i.i;;'i!l Class !>- \^ -^ G \j I Book ^^ Copyri^htN^. CDPmiGHT DEPosrr EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES lElmfatUmal Pagrlyologg ilottograylyH iEitUfli bg dug Mantca&e Wif'xpplt No. 7 Experimental Studies of Mental Defectives A Critique of the Binet-Simon Tests and a Contribution to the Psychology of Epilepsy ,. ■' By J. E. WALLACE WALLIN, Ph.D. Director of Psychological Clinic, School of Education, University of Pittsburgh Author of Optical Illusions of Reversible Perspective, Spelling Efficiency, Etc. laltlmnr?, 11 &. A. WARWICK & YORK, Inc. 1912 Copyright, 1912 By WARWICK AND YORK, INC. ©CIA. 3432 EDITOR'S PREFACE. In this, the seventh of the series of Educational Psy- chology Monographs, Dr. Wallin has presented the results of a systematic critical study of the Binet- Simon scale when applied to a colony of over three hundred epileptics. These results will appeal to schoolmen and to physicians and alienists for two reasons. In the first place, they have added to our knowledge of the men- tal status of the epileptic. Epilepsy has long re- mained a little-understood disease. We are told that some men who have attained eminence have been epileptic, e. g., Julius Caesar, Mohammed, and prob- ably St. Paul, to cite conspicuous instances ; yet the great majority of those afflicted with the malady fail to reach normal mental maturity, and fall, as Dr. "Wallin shows, into the upper group of the feeble- minded, now known as morons. These children re- semble more the typical laggard of the public schools than the typical feeble-minded child, and they re- quire special educational treatment. For this reason the more we can discover as to the psychology of the epileptic, the more successful will be our educational measures. In the second place, these results have added to our knowledge of the Binet-Simon tests. The fact that these tests are being widely adopted by school authorities as a convenient and assumedly scientific vi editor's preface method of measuring retardation or acceleration of mental development makes any careful study of their reliability of direct practical importance. And the more so, because, as Dr. Wallin's investigation shows, the tests in question are far from being so simple and so universally applicable a tool as many laymen have supposed. On the contrary, despite their undeniably great practical value, they suffer from numerous imperfections and limitations. These defects can be made known only by thoroughgoing trial on large groups of individuals by expert inves- tigators. This monograph makes a valuable contri- bution to this critique of the tests. Dr. Wallin be- lieves that the reconstruction and perfection of the tests by competent research should not be allowed to conflict with the continued use of the 1908 scale according to prescribed standard conditions. For this reason a guide to the conduct of the 1908 series (reprinted, with the permission of its editor, from The Psychological Clinic, December 15, 1911) is in- corporated herewith as the final chapter of the monograph. Guy Montrose Whipple. TO MY FIEST PRECEPTOR IN PSYCHOLOGY AND PHILOSOPHY, EDWARD FRY BARTHOLOMEW. FOEEWORD, The following experimental studies represent some of the fruits of my psychological and anthro- pometric investigations of epileptics in The New Jersey State Village for Epileptics at Skillman from October, 1910, to and including May, 1911. A second comparative psychological research of public school children and epileptic school children, by a set of serial rate tests of development, will appear in a separate volume at some later date. I feel that no apology is needed for the time and pains devoted to the critical examination of the Binet-Simon scale of intellectual development. This scale has recently been victimized by the indiscrimi- nate exploiter. It has been hailed by popular writers in the daily and periodical press, and even by scien- tific workers, as a wonderful mental X-ray machine, which will enable us to dissect the mental and moral mechanisms of any normal or abnormal individual. But these tests are no ''open sesame" to the hu- man mind, no talisman that will transform an ordi- nary observer into a psychic wizard. Because the scale is coming into wide use in the public schools, the psychopathic and criminological institutes, and institutions for mental and moral defectives, and because it is being appropriated by ordinary class- room teachers and persons having no technical train- ing in clinical psychology or knowing little about scientific method in general, it is worth while to point 2 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES out its legitimate uses as well as its limitations and present imperfections. These tests need to be safe- guarded from uncritical exploitation and mystifica- tion, and rescued from the educational fakers and medical quacks. I shall feel well repaid for my labors if this contribution arouses students to a full realization of the necessity of a further prolonged and systematic critical study of the scale. These studies — as well as those which will fol- low — should appeal to the students of defectives and the problems of human variation. The serviceability of the Binet-Simon scale, despite its imperfections, in determining the mental status of defectives, will be seen. Moreover, the study of epileptics should interest particularly the alienist and physician, and the schoolman: the alienist and physician, because epilepsy is a pathological condition as yet little un- derstood, characterized by marked mental disturb- ances; the schoolman, because the epileptic children constitute a numerous class which grades nearer the public school laggard than do the feeble-minded chil- dren, and which cannot be reached by the cut-and- dried methods of the schools, but which requires a special educational regime. I have designedly contented myself with a purely empirical and experimental exposition, leaving the reader to work out the implications and draw his own conclusions from the facts supplied. The facts tell their own story. To have pointed out the vari- ous practical and theoretical implications affecting the education, care and training of epileptics would have unduly extended the size of the monograph. J. E. W. W. February, 1912. CONTENTS. Editor's Preface v Foreword 1 CHAPTER I. Introduction. The Binet- Simon Scale as an In- strument for Classifying" Defectives .... 5 CHAPTER II. Testing the Binet-Simon Tests : An Exemplifi- cation of an Adequate Method of Analysis . 20 CHAPTER III. The Variation of Mental and Physical Traits in Relation to the Age Classification of the Binet- Simon Scale 59 CHAPTER IV. A Practical Guide for the Administration of the Binet-Simon Scale for Measuring Intel- ligence . 116 CHAPTER I. INTRODUCTION. THE BINET-SIMON SCALE AS AN INSTRU- MENT FOR CLASSIFYING MENTAL DEFECTIVES^ The functions of a clinical psychologist in an insti- tution for defectives, in a public school system, in a university, in a psychiatric institute, or in a juvenile court, are twofold: first, that of theoretical inves- tigation, or the increase of knowledge under con- trolled and verifiable conditions. This is essentially the work of the research psychologist or of pure science, so-called; second, that of practical applica- tion, or the utilization of the truths discovered in the educational, vocational, recreational, hygienic, med- ical and custodial treatment of the sufferers. This is the work of the consulting psychologist as distin- guished from the pure researcher, and constitutes the sphere of orthogenesis,^ mental hygiene, or ap- plied clinical psychology. While the line of demarca- tion between these two aims should not be made too fast and hard, logically the work of investigation in an infant science should take chronological prece- dence to the work of consultation, as, indeed, science logically must precede true art. The art of righting defectives cannot rise above the empirical until it is ^This chapter has been adapted, by permission, from a paper read before the St. Louis meeting of the National Association for the Study of Epilepsy, and printed in the Transactions of the Association for 1911, Vol. VIII, pp. 29f. ^In its twofold aspect of orthophrenics and orthosomatics, as ex- plained elsewhere : J. E. Wallace Wallin, Individual and Group Effi- ciency, Psychological Bulletin, 9 : 19X2, October. 5 C EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES based upon a foundation of assured facts. Until we thoroughly understand the different types of ner- vous and mental defectives our treatment cannot be made maximally effective. For these reasons I pur- posely aimed to confine my work in the psycho-clin- ical laboratory at Skillman entirely to research, and for these reasons I shall limit the discussion in the following pages almost entirely to a consideration and interpretation of the facts brought out in the investigations. During my eight months' stay at the institution the following surveys^ of the Village were completed (completed as far as measuring or testing each pa- tient once is concerned) : measurements of standing and sitting heights, of weight, of lung capacity, of the strength of right and left hand grip, of station or body sway, of the speed of performing the form- board test (replacing ten blocks of various forms in corresponding holes in a board), and of intel- lectual capacity, or the extent of intellectual retarda- tion, as evidenced by the Binet-Simon serial diag- nostic scale. In addition to the above, a special set of tests was carried out, designed to show the character and ex- tent of the deviation or disorganization found in epi- leptics in respect to a number of particular mental traits and capacities which play a basic role in men- 'The other lines of investigation started during this period comprise a series of measurements of the effects of convulsions upon various mental traits and capacities, and an investigation into the personal, social, motor, industrial and school efBciencies of epileptics. The lat- ter was carried out by meansi of the printed questionnaire or syllabus method. See Wallin, Human EflSciency : A Plan for the Observa- tional, Clinical and Experimental Study of the Personal, Social, In- dustrial, School and Intellectual Efficiencies of Normal and Abnormal Individuals, Pedagogical Seminary, 18 : 1911, 74ff. BINET-SIMON SCALE FOR CLASSIFYING DEFECTIVES 7 tal development. These tests include an extended series of controlled group measurements of the speed and accuracy of perception, perceptual dis- crimination, observation and reaction; the capacity to memorize, and the power of immediate and pro- longed retention; the capacity and rate of forming spontaneous associations with determinate ante- cedents ; the ability to form such controlled associa- tions as are involved in adding columns of ten one- place digits and of supplying antonyms to a set of simple words; the ability to retain a list of logical and illogical sequents, with determinate antecedents, from one reading by the experimenter both during a period of two minutes and during a period of four weeks ; the capacity for visual imagination, and the capacity for linguistic construction as evidenced by the ability to construct a maximal number of words from six supplied letters, and by the ability to form a maximal number of sentences each of which had to contain three supplied nouns or verbs. This set of differential tests was elaborated into six consecutive series and was given every twenty-eighth day to groups of the brightest epileptic school children at the Village and to somewhat less than 100 bright, average and backward pupils in the public schools of a nearby town. The results of the testing in the public schools will enable us to plot a series of nor- mal age norms of the functional capacity of the vari- ous traits tested. By means of such comparative scales or indices it will not only be possible to bring epileptics, or other types of defective individuals, into perspective with normal children, but to plot individual curves of developmental defects or accel- erations in the case of any abnormal or supernormal 8 EXPERIMENTAL STUDIES OP MENTAL DEFECTIVES child. Moreover, since all of the successive tests, while different, are practically of the same difficulty, it is possible, by giving them at stated intervals (monthly, semi-annually, annually, etc.), to plot rate curves of development ; that is, it is possible to meas- ure experimentally the rate of improvement or de- velopment which various mental traits undergo from time to time as a result of normal maturation, or of education, training, practice, or familiarity.^ The desirability, or even the feasibility, of estab- lishing psychological rate-norms of development has, strangely, scarcely dawned upon us until recently, although the practical value of such norms is pos- sibly greater than the value of the corresponding anthropometric standards of yearly development during the growth period of height, weight, physical energy and vital capacity. The importance of a set of anthropometric norms, arranged on the grade or percentile basis, has been eloquently set forth by the lamented Sir Francis Galton, to whose comprehen- sive intellect many sciences have become indebted. Thanks to the labors of a few of Galton 's followers, notably Bowditch, Porter and Smedley, we now pos- sess a set of fairly reliable norms of physical devel- opment for certain ages, by means of which we are able to determine the physical station of a given child of a given age, and by means of which we can say whether his physical progress is normal or sat- 'One of my aims in devising these tests was to secure experimental means for measuring results in education, such as the effects of dental and medical hygiene, schoolroom moisture and temperature, diurnal and seasonal changes, fatigue, the use of tobacco, upon the working efficiency of the pupil, and upon the growth and development of various mental functions. A description of five of these tests will be found in Wallin, Experimental Oral Euthenics, Dental Cosmos, 54 : 1912, 404- 413; 545-566. (See also references.) BINET-SIMON SCALE FOR CLASSIFYING DEFECTIVES U isfactory as measured by the percentile grade to which he belongs (using height as the basis of com- parison). But we are now beginning to realize that we cannot properly diagnose developmental defects of the mind until we have constructed a similar set of psychic norms of development of various traits and capacities. When we have psychic norms for specialized, capacities, we shall be able to locate the mental station of a given child at a given time, and determine whether his rate of mental development is normal for the grade in which he classifies. These norms will possess fundamental value for purposes of developmental diagnosis in the study of not merely the lesser deviations, but also the more pro- found mental abnormalities. To supply these men- tal developmental scales is chiefly a matter of time, labor and ingenuity : the instrumental and technical difficulties are secondary. Such scales will not, of course, attain the accuracy of refined physical meas- ures, but they will be far superior to our present 'common sense' judgments. The fair degree of suc- cess attained by the simple Binet-Simon tests of in- telligence justifies the belief that this problem, baf- fling as it seems, is not insoluble. By means of the above serial group tests I am hoping at some later date to make some little addition to our knowledge in this largely unexplored, but inviting and impor- tant, field of inquiry. Aside from the value which the data from these tests will have for developmental diagnosis, the results may also be used as a means by which to check up the accuracy of the Binet-Simon tests, with the consideration of which I shall be chiefly concerned in the present volume. The latter tests are such a striking contribution to our methods 10 EXPERIMENTAL STUDIES OP MENTAL DEFECTIVES of studying defectives that too much time cannot be spent in the effort to give them their proper eval- uation. Those who may be unfamiliar with the B.-S. (Binet-Simon) tests should consult Chapter IV. It may be explained here that they constitute an ex- tremely simple and yet fairly serviceable measuring rod of intellectual capacity. The scale consists of a graded series of 62 individual tests (including the one- and two-year-old tests), varying in number from three to eight for each of the first thirteen years of life. To illustrate: a child who follows visually a lighted match moved in front of his face, who grasps and handles a block placed in his hand, and who grasps a suspended cylinder, is credited with a men- tality of one year. A one-year-old child normally does these things. A child who can state his sex, who recognizes common objects, such as a knife, penny and key, who can repeat three numerals when heard once, and who can designate the longer or shorter of two lines differing by one centimeter, is rated as four years old mentally. By means of a scale of this sort it is possible, therefore, to classify individuals approximately according to their degree of intelligence, and to measure the extent of intel- lectual arrest, retardation or degeneration shown by subnormal individuals, or the extent of precocity or acceleration shown by supernormal individuals. This scale is the product of laborious and ingenious research by the noted French psychologist, Binet, and his co-worker, Simon. The standards fixed for each age are supposed to represent the normal per- formances of French children of the working classes. In the present chapter I shall limit myself to show- BINET-SIMON SCALE FOR CLASSIFYING DEFECTIVES 11 ing how a colony of epileptics classify by the B.-S. scale, and to pointing out certain striking peculiari- ties or anomalies in the curve of distribution, which, while they may be caused by various factors, either implicate fundamental abnormalities in the mental make-up of the epileptic or fundamental defects in the B.-S. tests. GRAPH I. Classification of 333 Epileptics ( ( ) by the Binet-Simon Method. ) and 378 Feeble-Minded B.-S. Ages 9 10 11 12 13 BINET-SIMON SCALE FOR CLASSIFYING DEFECTIVES 13 «.j vi & a cr P P a fB •-< B "^^ cr o • P CO o I— I S; MO CD t» M P 5 H-( Qp o ^"^N. pa* '-I c ^ <<] "^ tr l-h 5 C O O CJ* Si I P IK P £ ^^ wP c a ,B P M 5 B 00^ P 2 p o CO J? p ■' g B B o a- P to 2 sr til IX.' >-*■ i-v O hH W p ■•p B P-P'G. •^ 2- s^ '-' P 1 P* o M «P HH X o .►1 to CO M § b b o lO o 1*- bx to CI CO 1-1 CI CO 00 b CO -1 § bo bo -1 Oi CO 05 o CO CT ^1 M Oi Oi ;. CI 00 b CO lO CI 05 J(i. CI t-H hH K^ 1^ l-H B I-l ^ ^ Li hH hH hH ■< •< hH fh B.-S. Age. II M 1-1 t3 05 05 05 i^D CD cji >;>. h-1 ct *■ -1 to l-i to h-1 o CO 00 CO 00 CO 00 :-5 CI M ;_i QC CO 05 p CD to GO 05 00 o •-a CO bo M b CO b b ^ h-1 CO h-1 ~t. w fe; M to h-i to to to too CI c;i -^ o o CI CO -1 CO to > "^ t pi '~- to CO to CO CI to CD O to 05 h-1 to CI hfi. C3 o 05 -?slis to M to h-1 B- -, '^ to j)^ CD 00 00 ^ b b CO CI b h-1 rfs. CO hfi^ h-ib ^b CI CI I. -Simo over 2; en. 0.9 4.6 tOCI M hP^ CI to h-1 h-1 -^co?g. Z^ S 00 o ^ 50 o 1-1 h-1 to -1-1 1+^ o ^^^ ■ p" o to 1-1 to m ■~s CO 00 05 00 ^ to JO CI CO CO h-1 rf^ h-1 C0° ^ CO CI b t- CI b b "hP^ bo h-1 h-1 ^ i^ CO CO feg to >M_i to k^ h-1 h-1 h-1 05 t^p g p ■« -1 t-1 CO 05 CO ><^ hJ^ h-1 >^-100 h-1 l-i to M to •iH* CO M CO 00 CI .^ :^ p to CO *^ 05 coio^g co:■ to M 5* rf^rf^OOlOiCOrf^hl^COW. ^ S-l-100rf^K-l)-lOOO^S Ol^OWCOtf^OOO to 05 to rf^ to SOCOOC-ltOQOO U^0l^5 0®000 M to -q to *■ to OOOrf^C^tO>+^OCO -q-.ikt^i-'i-iooo CJIOOSOOOOOO (_i 1-1 tsJ -l to *>• l-i Ol-'M^>f^tOlOCH*'CO MM _ OOCDCOOOOQQ oo<:Jirf^o^ooo M M to CO MO to CO Ol W op CO *^ o o g <1 I— I O^ I— I o o 2 MO M O M M M to 1-1 HI CD ►<:>. CD CO Oi 05 O QQ Q CO i^ o o oo M O -1 --< en to ' O i o 8 CO too o o C CO CO o ~ o X M ^ 13 S- t- ^u 52 - ^ ? a 2 S ;5 ^ CD 00 Ol a ^ s a- "S a a •^ 00 a. Oo CIS a. a- a «. m >i S ^ =t '^ «-(. s- "SS 3 feg TO m TO "« a* a a f 00 (;i S ^ I a g S' TO a* s< 00 25 I— I ?^ 2- 30 EXPERIMENTAL STUDIES OP MENTAL DEFECTIVES The importance of this factor may be seen by re- ferring to Table III, which shows that in not more than five ages (III, IV, X, XII, XIII) did more than 60% of the patients successfully pass the ages in which they classify. The ages which make the worst showing are VI, VII and IX, particularly VI and IX, which proved veritable pontes asinorum. It follows that the age classification of the majority of the epi- leptics was attained on the basis of advance points. Very few, except in Age X, passed all but one of the tests of the mental age assigned them. It is therefore apparent that any B.-S. testing which is confined to the subject's chronological age or to the age immedi- ately above or below is of questionable value, if not worthless, and is positively pernicious to the inter- ests of scientific research either for purposes of diagnosis or classification or the testing of the scale. To what extent, then, is the wide-range testing responsible for our peculiar curve of distribution? Owing to the present inaccessibility of the original records, I cannot answer this question in exact numerical terms. But the two following considera- tions would seem to indicate that the method of test- ing is of secondary consequence. First, it is probable that other examiners (particu- larly the Vineland workers, with whose results our curve has been contrasted) have made a practice of testing the mentality of the examinees at least tivo years above the station in which they grade (since the basis of rating has been uniformly the highest age passed, the testing of lower ages can be entirely ignored). In case this is granted — other writers have been silent on this point — the fundamental question reduces merely to one of the relative TESTING THE BINET-SIMON TESTS 31 amount of the advance credits given — tlie difference in the number of credits which may be gained by testing only two years above the age fully passed or by testing in a considerable number of ages. While we have found numerous epileptics whose capacities have ranged throughout the larger part of the scale, the successes higher up have usually been sporadic and exceptional, so that it is entirely probable that the great majority of the advance credits given the epileptics have come from the two adjacent higher ages. If this is so, the peculiarities in the curve can- not be due primarily to the wide-range testing. Second, while the system of advance crediting from a wide-range testing effectually conceals the skew which should have appeared in the curve at Age VI, it is only very slightly responsible for the skew at Age X, because 94% of those who grade X pass the standard for this age, while 84% of them fail on the IX-year tests. Not only so; an examina- tion of the feeble-minded classification shows that the most obvious break occurs in the ninth standard. It is thus apparent that if further analysis reveals other significant factors, as implied in the foregoing statements, the methods of scoring and wide-range testing must be given a subordinate place. There are three methods which we may employ for a fur- ther critical examination of the data : Percentage of Epileptic Children Passing Each Individual Pitiet-Simon Teat, trrespectivc of the Age in tchiclt the Children nre Classified B.-S. , Boys V , Girls . Test. No. % Range. No. % Ilange. I. 1 11 100 7 85 ,-Aves. . Each Age.—, -Both Seies-v Both Avt. Ave. Boys. Gills. Sexes. No. % Range. % % % 60 91 58 42 97 33 102 94 23 69 18 83 41 75 23 CI 19 68 42 «!<; 29 83 2S 71 57 77 20 75 IS 19 68 15 39 71 2.1 04 25 68 .50 (^i 21 62 24 70 45 66 .'52 60 37 r.i 89 .56 43 8.S 28 35 91 40 78 s:i 2i; 6.5 30 70 .56 67 43 14 38 10 81 10 27 60 31 58 .5.H .59 47 79 36 63 S3 72 27 .-!(; 29 51 .56 43 .-)4 44 .3.S 36 92 41 30 60 0.5 32 68 60 62 64 .30 61 35 77 74 68 3G 7.5 :!0 66 66 71 28 tlS 28 71 56 69 4.5 66 34 44 79 57 47 42 40 42 87 42 VIII, 35 47 7J 3B 47 61 i 37 51 57 38 39 51 39 45 51 40 38 73 IX. 41 47 64 42 45 82 43 44 43 44 45 17 45 47 34 XI. 51 33 52 33 53 39 54 33 29 XII. 56 31 57 M .58 33 59 30 XIII . 60 29 61 29 34 14 73 36 21 52 30 33 54 40 V.i 65 13 23 .50 16 Ave 39 38 37 B.-R. Test ; the tests are niiinlwred consecutively, following the customary order (See Form I of the Skillman blanks). No.=i:aimber tested. %=l>er cent, of successes. R.ince^difference between the highest and lowest per cents, in eath age. Aves. for each age represent the sum of all the tests in each age, divided by the iiuuiher of tests. The above includes all the epileptic chiMieu except nine. It includes four boys who had been free from attacks for over two years. TESTING THE BINET-SIMON TESTS 33 (1) We may determine the percentage of suc- cesses (passing) for each individual test of the entire scale for a given group of examinees, irrespective of the ages in which the examinees classify. Such a tabulation for all the epileptic children (all under 21) appears in Table IV. The records for the adults are not now available, but the curves for the children and the adults, in spite of various interesting differ- ences, have the same general characteristics (Table I). As will be seen from Columns 2 and 5, Table IV, the number of children put through each test differs considerably, but the percentages of successes are in all cases based upon the number examined for the given test. (2) We may determine the per cent, of successes in each individual test, based only on the records of those subjects who classify in the age to which the test has been assigned. That is to say, we determine what percentage of patients who grade VI pass each one of the Vl-year tests, what per cent, of the VII- year-olds pass each of the eight tests of that age, etc. With this object in view, the data for the children have been retabulated in Table V. If the tests in each age are relatively equi-difficult, the successes should be approximately equal for the different tests of the same age. This method could be extended so as to include a determination of the per cent, of suc- cesses for each test in both loiver and higher ages. What per cent, of subjects who grade V, VI or VII pass each of the tests in Age III or IV? 35... 36... 37... 38... 39... 40... Ave. 33 71 66 26 66 85 66 85 LOO 100 34 EXPERIMENTAL STUDIES OP MENTAL DEFECTIVES TABLE V. Percentages of Children Passinc; the Individual Tests in the Binet-Simon Ages in which they are Graded. AGES I AND II. Both AGE III. Both , — Boys — ^ , — Girls — , Sexes. , — Boys — s , — Girls — ^ Sexes. Test. No. % M.V.No. % M.V. % M.V. Test. No. % M.V.No. % M.V. % M.V. 1.... 7 100 2 100 100 7.... 4 100 2.... 100 100 100 8.... 00 3.... 83 50 75 9.... 100 4.... 43 50 58 10.... 100 5.... 28 00 22 11.... 100 6 28 50 33 Ave..' 63 .30 58 .27 64 .27 Ave.. 80 .32 66 .13 73 .20 AGE IV. AGE V. 12.... 5 100 4 100 100 16.... 1 100 2 50 66 13.... 100 75 88 17.... 00 100 66 14.... 60 75 66 18.... 00 50 33 15.... 80 25 55 19.... 100 50 66 Ave.. 85 .15 68 .22 77 .17 Ave.. 50 .50 62 .17 58 .12 AGE VI. AGE VII. 20.... 4 50 6 66 60 27.... 5 80 6 100 90 21.... 00 00 00 28.... 60 100 82 22.... 75 50 60 29.... 100 100 100 23.... 50 50 50 30.... 40 33 36 24.... 25 66 50 31.... 40 16 27 25.... 25 5 00 10 32.... 60 33 45 26.... 75 66 70 33.... 80 100 90 34.... 80 66 72 Ave.. 42 .16 42 .12 42 .13 Ave.. 67 .17 68 .31 67 .24 AGE VIIL AGE IX. 66 6 66 66 41 6 100 3 66 88 55 50 53 42.... 100 100 100 44 100 66 43.... 50 33 44 55 33 46 44.... 16 00 11 11 66 33 45.... 33 2 00 22 77 66 73 46.... 33 66 44 51 .16 63 .14 56 .12 Ave.. 55 .29 44 .33 51 .28 AGE X. AGE XI. 47.... 16 93 8 87 88 51.... 6 50 2 50 50 48.... 100 75 91 52.... 66 100 75 49.... 87 75 83 53.... 33 50 37 50.... 68 87 75 54.... 33 50 37 55.... 33 00 25 Ave.. 87 .09 81 .06 84 .05 Ave.. 43 .12 50 .20 45 .14 AGE XII. AGE XIII. 56.... 2 50 1 100 66 60.... 100 57.... 50 100 66 '61.... 100 58 50 00 33 62.... 50 59 50 100 66 Ave.. 50 .00 75 .37 58 .12 Ave.. 83 .22 Test=number of individual B.-S. test. No.=numl)er of patients examined. M. V.=mean variation (between the different subjects' scores based on the figures given in the per cent, columns). Average M. V."s for all ages, except XIII: boys, .20; girls, .21; both sexes, .17. Average per cents, for all ages, except XIII : boys, .61 ; girls, .61. TESTING THE BINET-SIMON TESTS 35 (3) We may determine what percentage of pa- tients classifying in a given higher age, say VIII, IX or X, fail to pass any of the loiver age-norms (not individual tests, but the age standards) . For exam- ple, how many X-year-oIds do not pass Age VI? This throws light on the collective difficulty of the tests of various ages. For this study we have two sets of data, one incomplete and the other complete. The incomplete data for 276 juvenile and adult epileptics. Table III, were originally gathered for another pur- pose, namely, the determination of the per cent, of patients who passed ages immediately adjacent to that in which they classified. No record was made of the ages in which they did not pass any tests at all ; however, for some of the patients the data cover more than the years contemplated. This explains why the per cents, in some cases are based on many subjects, but in other cases on only a few. More- over, data based on only a few cases have not always been tabulated. In the absence of the original rec- ords, these defects cannot now be remedied. But, in spite of these defects, a comparison of the ages in Tables VI and III, which are comparable, indicates that the data are sufficiently reliable for our pur- pose. In Table VI we have in complete form for the entire population of the institution the percentage of patients between the B.-S. Ages of IX and XIII who pass Age IX and the per cent, of those between Ages VI and X who pass Age VI. What, now, does a critical examination of our tables reveal ? 36 EXPERIMENTAL STUDIES OP MENTAL DEFECTIVES TABLE VI. Percentage of Epileptics Grading from IX to XIII who pass the B.-S. Standard for Age IX. Entire Both Popu- B.-S. Boys. Girls. Sexes. Men. Women. Adults. Males. Females, lation. Age. No. % No. % No. % No. % No. % No. % No. % No. % No. % IX 6 16 3 33 9 22 10 10 9 00 19 05 16 12 12 08 28 10 X 16 37 8 00 24 25 27 15 32 12 59 13 43 23 40 10 83 16 XI 6 82 2 00 8 62 15 20 5 00 20 15 21 38 7 00 28 28 XII 2 100 1 lOO 3 100 5 60 3 66 8 62 7 71 4 75 11 72 XIII 2 100 50 9 66 7 85 16 75 11 72 7 85 18 78 Ave. 67 33 51 34 32 34 43 231 40 percentage o f Epileptics Grading from VI to X who fnr Anp. VT pass the B. -S. I Srtan idar VI 4 00 6 00 10 00 7 00 5 00 12 00 11 00 11 00 22 00 VII 5 00 6 33 11 18 9 00 12 00 21 00 14 00 18 11 32 06 VIII 9 44 6 33 15 40 20 10 20 25 40 17 29 20 26 26 55 27 IX 5 60 3 66 8 62 10 70 9 44 19 58 15 66 12 38 27 50 X 14 78 8 87 22 81 23 65 31 48 54 55 37 70 39 56 76 63 Ave. 36 43 40 29 23 26 31 26 29 Age, the age in which the patients grade. No., number of patients in- cluded in the average (including those who pass and fail). %, per cent. passing. This includes all who did not fail in more than one test in each age. ^Exclusive of 85% in Age XIII. In Table IV we find a surprisingly large variation in the difficulty of the individual tests, both for the boys and the girls, ranging from 100% (first test, Age I, boys) to 0% of successes. Since this variation is between tests occurring in any part of the scale, it is significant only because of its extreme character, for a large variation will inevitably occur when sub- jects of very varying capacities are tested through- out a large part of a graded scale and the results are thrown together. The more stupid will necessarily fail on all the higher tests, and the brighter will suc- ceed on all the lower ones. Hence, all that could be demanded, at the utmost, is that the lower-grade subjects find the high-grade tests in the same ages about equally difficult, and the high-grade subjects the tests in the same lower ages about equally easy. Accordingly, we are justified in comparing only the tests of the same ages. Assuming that these are fairly uniform in difficulty, the percentage of fail- TESTING THE BINBT-SIMON TESTS 37 Tires should be approximately the same for all the tests of the same age. It is patent from a cursory glance at Table IV, however, that there is a wide difference between the tests. This is seen most rapidly by comparing the ''range" columns. The range between the highest and lowest per cent, of successes in each age aver- ages 37% for the thirteen years. The general aver- ages are practically the same for the boys and the girls, although there are striking differences for some of the ages. The largest range is between the tests of Age VI, 62% ; XII, 57%, and IX, 56% ; and the smallest between Ages IV, 11% ; VIII, 21%o, and I, 24%. The order differs somewhat, however, for the boys and the girls, the largest range for the boys being in Ages VI, IX and XII, and for the girls in II, VI and XII, in the order given. The smallest are in Ages IV, I and VIII for the boys, and X, IV and XIII for the girls. If we now regard the collective (average) difficulty of tests of the same ages, we again find a very con- siderable variation. The most difficult ages are XIII (14% of successes), XII (21%), XI (22%) and IX (37%), and the easiest I (82%), III (76%), IV (72%) and V (69%). The order is the same for the boys and girls considered separately, except that II displaces IX for the boys in the difficult set, and III and IV exchange places in the easy series. The aver- ages for the boys and girls are not appreciably dif- ferent, except in Ages I, III, VIII, IX and X. It is evident, however, from what has been said that averaging the per cent, of successes by ages in this fashion (last three columns of Table IV) is val- uable chiefly for purposes of comparing the relative 38 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES efficiencies of the sexes in the same ages, and not primarily for the purpose of comparing the diffi- culty of the various ages. The method is faulty for the latter purpose, particularly as affects the ex- treme ages in the scale, because, since the patients were tested throughout a large extent of the scale, some comparatively high-grade subjects were tested on the lower tests, while, vice versa, some low-grade subjects were tested in the higher ages. The tend- ency, therefore, is to decrease the real difficulty of the lower ages and to increase that of the upper ages. This, indeed, the results show. If, therefore, we confine the comparison to the middle range of the scale, V to IX, inclusive, where these tendencies may neutralize one another, we find that Ages IX and VI are the most difficult, followed hy VIII, VII and V (decreasing order). This conforms substantially with the data in Tables V and VII, to which we now turn. The latter enable us to make a more reliable com- parison between the difficulties of the different tests of the same age, and indirectly between the relative collective difficidty of the different ages, for here, as explained, the per cents, of successes for a given test are based on the testing of those subjects only who classify in the age in winch the test has been placed. The variation between the different tests of the same age appears most readily from an examination of the M. V. and "range" columns. The average M. V. for all ages amounts to .17, a very considerable fraction of the size of the average per cents. (.61). It is practically the same for the boys and the girls, although there is considerable difference in a few of the ages. TESTING THE BINET-SIMON TESTS 39 TABLE VII. Ranges 'between the Highest and the Lowest per cents, of successes recorded for the tests of each Age in Table V. Ages I & II. III. IV. V. VI. VII. VIII. IX. X. XI. XII. XIII. Boys 72 100 40 100 72 60 66 84 32 33 00 50 Girls 100 67 72 50 66 84 67 100 12 100 100 Children . 78 74 45 33 70 73 40 89 16 50 33 The general averages for all Ages, except XIII, are as follows : boys, 60% ; girls, 74% ; both sexes, 54%. The M. V. amounts to more than .25 in seven cases : Ages I-II, boys and girls; III, boys; VII, girls; IX, boys and girls; and XII, girls (exclusive of Age V, boys). The ranges (Table VII) are also very large for these ages, the seven largest ranges being in Ages I-II, boys ; III, boys ; VII, girls ; IX, boys and girls; XI, girls; and XII, girls. • Five of these amount to 1.00 and two to .84. The average of the ranges for the years I to XII, inclusive, reaches the very considerable sum of .54. It is somewhat larger for the girls than the boys. On the other hand, considering the minimal varia- tions, we find the M. V, to be less than .17 in the fol- lowing ages : III, girls ; IV, boys ; VI, boys and girls ; VIII, boys and girls; X, boys and girls; XII, boys. In some of these ages we also find, correspondingly, the smallest ranges, namely. Ages IV, boys ; X, boys and girls; XI, boys; XII, boys and girls. The smallest range comes in Age X. If we consider only the general averages (the av- erages for the two sexes), the largest M. V.'s (from .20 to .28) are in the following ages: I-II, III, VII and IX, and the largest ranges in Ages I-II, III, VII and IX. The smallest M. V.'s (from .14 to .05) are in the following ages : V, A^E, VHI, X, XI and XII, 40 EXPERIMENTAL STUDIES OP MENTAL DEFECTIVES and the smallest ranges in the following: V, VIII, X and XII. Summarizing these results, it appears that, if we exclude Ages I-II as representing tivo years rather than one, the greatest variation occurs between the tests of Ages III, VII, and particularly IX, and that the greatest uniformity obtains in VIII, X (dis- tinctly), XI and XII (less decidedly in IV and VI). The question now naturally arises whether the collective difficulty (average per cent, of successes for all the tests of a given age) is greatest for the most variable age-norms. Reference may be made to three tables to obtain light on this question. On the basis of the average per cents, in Table V, there appears to be no correlation. The group average for Ages III, VII and IX is .64 (.66 for boys and .59 for girls), and for Ages X, XI, XII, IV and VI .61 (prac- tically the same for the boys and the girls). The per cents, in Table V, however, do not represent the number of patients who passed the age-norms, as in Table III. In the latter table the successes (32%) for the group of variable age-norms are appreciably less than for the group with more uniform tests (49%). It is noteworthy that the successes are above 63% in three ages (IV, X and XII) where the varia- tion is small, and in only one age where the variation is large (Age III), and that the successes are sur- prisingly small in the remaining two ages having unequal tests (VII and IX. Cf. also Table VI). In the case of Age VI, however, the variation is small, although this is a very difficult age (Tables III and VI). It is therefore apparent that there is a greater variation between the tests of a difficult than an easy TESTING THE BINET-SIMON TESTS 41 age-standard, although this need not always be so. If the tests of a given age are uniformly difficult, the variation will be small, and the same is true, of course, if they are uniformly easy. But they are more liable to be uniformly easy than uniformly difficult. The most striking result which our analysis of the M. V.'s and ranges for the ages has thus far shown is the amazing lack of uniformity between the diffi- culty of the tests of the same age-norms for fully half of the ages of the scale — amazing from the standpoint of the precision demanded by the stand- ards of scientific work. Moreover, we have also seen that the collective difficulty of different age-stand- ards differs — to how great an extent may be seen by a re-examination of Tables III and VI. The per cents, of passing range from .00 to 100%. (The lat- ter may be ignored, as it is for the final age in the scale, and all who passed this standard could only do so by virtue of the tests of this standard itself). The successes exceed 70% in only four ages (III, IV, X and Xin), and reach less than 42% in six ages (V, VI, VII, VIII, IX and XI). The most difficult age-norms are VI (00% of passing), IX (10%) and VII (13%); and the easiest III (.73%), X (84%), XII (58%), and XIII (100%). The last age may be neglected for the reason given, and III and XII be- cause of the fewness of the subject tested in these ages. We thus obtain for epileptics one specially easy age-standard, X, and several specially difficult age-standards. In order to study more thoroughly the two most difficult ages, VI and IX, the data for the juvenile and adult epileptics were retabulated as appears in Table VI (which see). 42 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES With this surprisingly wide range in the difficulty of the Binet-Simon age-norms demonstrated, we are forced to meet a fundamental question, namely : In order that a measuring scale of intelligence shall approximate the character of scientific measures, what amount of variation in the difficulty of the tests of the same age shall he considered as the maximal permissible? What percentage of normal children should pass the tests designated as criteria of their chronological ages in order that the tests shall con- stitute scientific standards for these ages? It is evi- dent that if only a small per cent, of typical children pass the standard tests the latter are worthless. It is equally evident that 100% of passing is also out of the question, because normal children will differ considerably in various traits and capacities, some being strong in one trait and some strong in another. We may thus regard as perfectly normal a certain amount of variation in the capacities of children of the same ages and training. Moreover, mental meas- urements are variables, and not absolute constants. Consequently the standards of each age must consist of a number of tests, sufficiently comprehensive to survey a variety of fundamental mental traits and capacities. And the difficulty of these tests (and the scoring) must be so adjusted that, while one test may be too hard for one child and another too easy, the majority of children should be able to pass the collective standard, i. e., their age-norm. To be more specific, I should hold that if 75% of normal (so- called) children fail to pass their age-norms, the norms are too difficult. Some of the tests must be transposed or eliminated. On the basis of this standard, it appears that only TESTING THE BINET-SIMON TESTS 43 two age-norms are properly constructed, so far as the testing of the epileptics is concerned, namely, III and X (Age IV falls short by 4%). Half of the ages do not even approximately satisfy the require- ments. In fact, we may lay down this rule: that epileptics do not qualify for a given age on the basis of satisfying the requirements of that age, hut on the basis of the system of advance credits from higher ages. ' To what, then, are the demonstrably large varia- tions in the difficulty of the various Binet- Simon age- norms due I Are they due primarily to inherent de- fects in, and misplacements of, the tests themselves — defective horizontal and vertical arrangement — or to fundamental deviations or abnormalities in the epileptic mind? To answer this question satisfac- torily we must still ascertain what particular tests produced the greatest 'mortality ' — ^were too difficult — and what particular tests were too easy. More than that, we must determine whether the tests found too difficult or too easy for epileptics have likewise been found too difficult or too easy for normal children or normal adults. A comparison of Tables IV and V (omitting Ages I, II and XIII, on account of the insufficiency of the data) indicates that the following tests were dis- tinctly too difficult for epileptics : Age V, arranging triangular pieces into a rec- tangle;^ VI, repeating 16 syllables,^ executing three commissions, and knowledge of own age; VII, re- peating five numbers, drawing a diamond, and de- *The success for this test is high in Table IV, because it was given to an unusually large number of high-grade subjects. "Namely : "We get up in the morning, eat, work and play, and then go to bed." (Variant forms infrequently employed.) 44 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES scribing actions ; VIII, counting backward and copy- ing from dictation ; IX, descriptive or classificatory definition, six memories, giving correct change (25c. — 9, or 8 or 7c.), and arranging six weights ; XI, arranging shuflfled words into a sentence, essential ideas in abstract definitions, and uttering 60 words ; XII, repeating 26 syllables,^ and inferring facts from given circumstances ; and XIII, all tests. In all ex- cept one of the above tests the per cent, of successes is less than 50 ; in most of them decidedly less. The failures for the higher ages are not brought out properly in the tables because of the form of the tab- ulation and because of the fewness of the subjects tested. While the tables are not so constructed as to indi- cate which tests are too easy (passed in lower ages), I came to feel that the following should probably be so regarded : Age V, counting 4 pennies ; VII, num- ber of fingers on hands, and counting 13 pennies ; IX, naming days of week in order ; X, naming moneys ; and XII, giving rhymes. Since the above was written, the results of Kath- erine Johnston's,- Goddard's^ and Bobertag's* test- ing of public-school children have, opportunely, ap- ^Namely : "The other day I saw in the street a pretty dog who car- ried in his mouth a basket of strawberries." (No variants used.) ^Katherine L. Johnston, M. Binet's Method for the Measurement of Intelligence. — Some Results, The Journal of Experimental Pedagogy, 1: 1911, 24fif. "Henry Herbert Goddard, Two Thousand Normal Children Meas- ured by the Binet Measuring Scale of Intelligence, Pedagogical Semi- nary, 18 : 1911, p. 232f. *Otto Bobertag, Ueber Intelligenzpriifungen (nach der Methode von Binet und Simon), Zeitsch. fiir angewandte Psychologie, 5: 1911, 105fi£. TESTING THE BINET-SIMON TESTS 45 peared. These, with other results already available,* will better enable us to explain the inequalities which we have found in the scale. Miss Johnston tested a mixed group of 193 ele- mentary, and 25 high- school girls in Sheffield, Eng- land; Goddard's assistants tested the entire school population (1547) of the first six grades of a New Jersey school system, and Bobertag tested 435 regu- lar and special pupils in the schools of Breslau. The former followed essentially Binet's latest (1911) revision, while Goddard, Bobertag (with certain ex- ceptions) and I followed the 1908 series. All three of these studies confirm some of my sus- picions with respect to the accuracy of the scale. Johnston refers to one girl who did 15 and another 23 tests superior to the age in which she classified, but she does not give averages, nor any intimation as to how frequently this happened. It is, as we have seen, a frequent phenomenon among epilep- tics. We need to know to what extent it appears with normals before we can generalize unequivocally about abnormals. She, too, found pupils who, unable to satisfy an inferior age standard, passed higher ^The following studies or discussions have appeared too late to re- ceive notice in the present discussion : J. C. Bell. Recent Literature on the Binet Tests, The Journal of Educational Psychology, 3 : 1912, lOlfif. Lewis M. Terman and H. G. Childs. A Tentative Revision and Extension of the Binet-Simon Measuring Scale of Intelligence, The Journal of Ednicaiional Psychology, 3 : 1912, 61ff, 133£e, 198ff. Edmund B. Huey. The Present Status of the Binet Scale of Tests for the Measurement of Intelligence, Psychological Bulletin, 9 : 1912, 160 (a review of the literature). F. Kuhlmann. The Present Status of the Binet and Simon Testa of the Intelligence of Children, Journal of Psycho-Asthenics, 16 : 1912, No. 3. An abridged discussion of the relevancy of the scale will be found in : J. E. Wallace Wallin. The Present Status of the Binet-Simon Graded Tests of Intelligence, The Alienist and Neurologist, 33 : 1912, May. 46 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES ones, but no figures are given to show how frequently this happened. She also found obvious inequalities in the difficulty of the age-norms. Her records show that of 30 9-year-old girls only one satisfied the IX- year norm, as against 24 who failed and 5 who passed Age X. Of 41 7-year-olds, 6 passed their standard, as against 26 who failed and 9 who passed Age VIII. Of 22 8-year-olds, 7 passed, 15 failed and 6 ''went above" VIII. Of 38 10-year-olds, 12 passed and 26 failed. Of 24 12-year-olds, 5 pased, 18 failed and one passed Age XV. These figures indicate that there is a wide divergence in the difficulty of the tests of the same age, and that, with normal {sic) English girls, the standards are too difficult, as a rule: there are more who fail than reach their age- norms. The most difficult age-norm, according to Miss Jolmston, was Age IX. These conclusions, it must be remembered, refer to the latest, or 1911, ar- rangement of the B.-S. tests. An examination of Goddard's table (Table I) shows that there were more 6-year-olds who could satisfy the VII- than the Vl-year norms (69 against 48) ; that a larger number of 8-year-olds stayed in Age VII than passed Age VIII (87 vs. 86) ; that very few of the 8-year-olds were able to satisfy the IX-year tests (only 16, as against 86 for the VIII- year norms), but there were actually more 9-year- olds who could pass the X- than the IX-year norms (58 compared with 56) ; that an unusually large number of 10-year-olds qualified for their standard, while an appreciably smaller percentage of 11-year- olds qualified for the Xl-year standard; and that more 12-year-olds classified as X than as XII (42 vs. 39). The results of both of these studies of pub- TESTING THE BINET-SIMON TESTS 47 lie school children (Bobertag gives results only for the individual tests) accordingly confirm my find- ings, based only on the results of those epileptics who grade in the various Binet-Simon ages consid- ered, particularly in respect to the disproportionate difficulty of Ages VI and IX and the dispropor- tionate ease of Age X (Miss Johnston's data, it should be said, to avert misapprehension, are based on too few cases for Age VI). It is obvious that we must conclude that there is a faulty vertical arrange- ment of the tests in the scale as at present consti- tuted. No only so : our findings in respect to the individ- ual tests are also confirmed, at least in part, as shown in Table VIII for the tests which I came to view with most suspicion. In this table the conclusions of Decroly and Degand, Binet's 1911 revision, and Johnston's, Goddard's, Bobertag 's and my own find- ings are compared. The Skillman figures are based on Table V. Table VIII may thus advantageously be studied with respect to our fundamental inquiry : Does the testing of normal (so-called) children show that our anomalous results are due to defects in the construction of the scale or to the peculiar mental organization of the epileptic? A study of Table VIII shows, in harmony with our findings, that the folloiving tests (with certain obvi- ous discrepancies) are too difficult: Age V: rear- ranging triangles, by G. (Goddard) and Bo. (Bober- tag). B. (Binet) retains the test, however, in Age V. VE: repeating 16 syllables, by G. (15% of successes). B. omits the test. VII : repetition of five numbers, by J. (Johnston) and B., who elevates the test to Age VIII, as does G., though his own per cent, of sue- 48 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES cesses in VII (74%) indicates that it is about right where it is. Bo. finds it properly placed. VIII: copying dictation, by G.; B., sustaining D.-D.'s (De- croly and Degand) objection (training test), omits it, but Bo. finds it rightly placed. IX : giving correct change, by G. and J., but not by Bo. (who used 100 — 20 = 80). Definition by description, by J. and G. Arranging six weights, by J. and Bo., and ap- parently by B., who places it in X, while D.-D. would place it in V or VI. XII : repetition of 26 syllables, by J., G. and B., but too easy by Bo. if the sentence is simple and meaningful. XIII : all tests, by all. TESTING THE BINET-SIMON TESTS 49 O (m o ■< p ^ Oi o « a M M M £"• t"' 7^(15 00 1» CO >^ > ^ (B 5 P 50 -(t> trap ^ Oi OQ p ^- , - fD O p O 1 B M "I -^ w ) B I-) re ■" ►! ! re < g.• -i«Sapg|g„gg Pi re t objWk-. DoSph>— Ep ^ESg'BSKg.g.BdH re B 2 '-'^ re S:' CTnro .t^ ^ 2 foorq 3 : a P9 SHW 2 &SS p p re ;^Bi ^ S" So? J9 re • oj re " "" '" » H-. aj rfk B p" h3h3H i-i >-( O P P © CP) o 2. B p mm 0° re 2 W P , fdfSOWOKKOOaiOOMWOOOWOwv! ppi-spBPPBBP-ii-jPiai-jHBpaQPP „. C. OS-BQaBBQOQBE:.oBBC.. Bbb a)»)«m;rBBH"'--B_»-BM^'-'-a!S:BBtH Q.&^Q-» fpQ j^w ft^rere&re 2 an QQi rt-Qja&_ta ^^ aa j^ are rt-rt-art-a o o _. o • CO ^o ^ <1<1 GO « wnj re i" mo p-p 9 l-l '2 3- B.B Loo ^boB ©to OI-'t«5C«p Sag^°2^&;88§iga o o o o o o • .0.-»*0'.00'..0.».0. -q-jre rf^MMrf^re coosre re cni^osf ooS* w o o o . o . ft *< H O p o ^.-jgoicocog-ah-c'reg re re „ tc B " CO rt-B So" t»'Std re re i-j i©S^B3- re • » o ^" g s.» 50 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES The folloiving tests, contrariivise, proved to he too easy, in liarmony with my results : VII : counting 13 pennies, by B., Bo. and Gr. D.-D. object to the test as too mechanical. VIII : naming four colors, by D.-D., B. and G. (Our VII- and Vlll-year-old epileptics required 7.7 seconds, and IX-year-olds 5 seconds, to name the colors.) X: naming money, by Bo., B., who has dropped it to IX, and perhaps by G., who finds 70% of successes in IX, but retains it in X. J., however, found it too hard. XII : three rhymes, by G., but not by Bo., nor by B., who has raised it to XV ( !) . Counting four pennies in V, which appeared to Bo. and myself too easy, G. and B. found about right, while D.-D. would discard it as a training test. The following tests, which I found too hard, would appear to he properly placed: VII: describing ac- tions, by B. and G. VIII : counting from 20 to 0, by B., Bo. and G. (considered a training test by D.-D.). IX: 6 memories by Bo. and G., but too hard by J. B. omits it on D.-D.'s objection (training). (Our IX-year-old epileptics average 4.8 memories ; our X- year-olds, 5.3, and our Xl-year-olds, 6.5). In respect to the following tests the discrepancies are more patent than in any of the above tests : VI : three commissions, found to be properly placed by G., too easy by D.-D. and Bo., and too hard by B. and myself. VI : giving own age, distinctly too hard for epileptics, considered valueless by Bo., omitted by B., and considered a training test by D.-D. XI : ut- tering 60 words, too easy by D.-D., too hard by Bo., right for Age XII by J. and B., and for XI by G., and too hard for epileptics (about right for Age XII, where the average number of words was 59.9). XI: abstract definitions (justice, charity, kindness, used TESTING THE BINET-SIMON TESTS 51 at Skillman), right for Age XII by B. and Bo., too hard for XII by J., slightly too hard for XI by G., and distinctly too hard for XI for epileptics. XI: rearranging words, right for XII by B. and Bo., and for XI by G., too hard for XII by J., and distinctly too hard for epileptics of Age XI. Having thus surveyed the available facts, we are now in a position to formulate a number of impor- tant conclusions. (1) The marked irregularities in our curve of intelligence for epileptics are partly explained by the system of advance scoring from a method of wide-range testing, and by the method of combining the records of a large number of adult epileptics and a smaller number of juvenile epileptics (the propor- tion of the adults among the epileptics being larger than among the feeble-minded, as we have already seen). It might be assumed that the latter circum- stance — to which we have merely adverted in the foregoing pages — would render the curve more or less irregular, for a defective child and a defective adult whose intellectual strength or capacity are the same, would not necessarily test out exactly the same by the B.-S. scale as at present constituted. We have, indeed, already noted here and there various differences between the children and the adults (others will be noted in the following chapter), and have found evidences of the loss of lower-age capaci- ties among the adults. To measure the influence of this factor, however, it is necessary to examine the table of distribution (Table I), and particularly the table which shows the number of higher-grade sub- 52 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES jects who pass especially the crucial ages, Ages VI and IX (Table VI). In the first table it is seen that, while there are about 3% more children than adults grading VI years, the number of IX-year-olds is approximately the same. The number of imbeciles among the chil- dren, however, is considerably higher than among the adults, while the number of morons and XIII- year-olds is noticeably less. In Table VI it is seen that of those testing from IX to XIII years, 22% of the children pass the IX-year tests, but only 5% of the adults (the differences for Ages X, XI, XII and XIII are also large, due possibly to the small num- ber of patients in some of these ages), while of those testing from VI to X, none of the children passes Age VI, as against 11% of the adults (with large differences also for Ages VII to X). While it is thus evident that an adventitious factor of this character — the averaging of the gradings of defective children and adults — will distort the sym- metry of the frequency curve, an examination of the separate columns for children and adults in Table I shows that skews are present in the classifications of each, particularly at Ages IX and X, only less prominently for the children than for the adults. Hence, we appear to be justified in the conclusion that the distortions in our curve are not primarily caused by the method of advance scoring from wide- range testing or by the merging of the grades of young and old defectives. (2) On the other hand, one of the significant causes of our skewed curve (skewed certainly as compared with the curve for feeble-mindedness) is the intrinsic defectiveness of the measuring scale, TESTING THE BINET-SIMON TESTS 53 for our analysis has demonstrated that there is a greater discrepancy in the B.-S. 1908 scale (and 1911 scale so far as Johnston's results indicate) than has hitherto been conceded or suspected. Its obvious inequalities affect not only some of the individual tests in various levels, but also the age-standards. Just how much of the irregularities can be ascribed to the defects in the scale it is impossible to deter- mine with definiteness until a re-examination of epi- leptics has been made by a scale rendered maximally correct and standardized for 'Hypical" American children. (3) It is, however, not only probable, but reason- ably certain, that the combined influences of the fore- going factors does not suffice fully to explain our curve. Three affirmative reasons may be given why another factor, namely, the peculiar mental make-up of the epileptic, must be considered. First, an in- spection of Table VIII shows that the percentage of failures on the difficult tests is abnormally large for the epileptics, in nearly every case very much larger than for the normal groups, so that the correspond- ing traits in the epileptics have at least suffered marked impairment. Second, some of the tests ab- normally difficult for the epileptics do not always prove difficult for the normals. Third, failures on some of the tests through various levels imply an impairment of the same or related functions. Thus, the epileptics suffer from a fundamental impairment of memory, as shown by consistent failures in all the tests of memory span (sentence and number tests), by the inability to reproduce six units from reading a short passage once, by the inability to recall their ages, and to remember and execute three simple com- 54 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES missions. They are also markedly retarded or feeble in the higher thought processes, the rational functions brought into play in the more intellectual tests and in the tests calling for an adjustment to new situations. This is shown by the incapacity to form a rectangle from two triangular pieces, to exe- cute a triple order, to construct an intelligible sen- tence from displaced words, to give descriptive or classificatory definitions of common objects or ab- stract definitions of simple qualities, or to infer facts from given situations (the situations employed in the test, it may be conceded, do not well fit American conditions), or to utter 60 words in three minutes. The epileptics suffer from a pronounced retardation of rate in the stream of thought and of motor re- sponse. They apparently also suffer from a blunt- ing of the kinesthetic sensitivity: the threshold of sensory discrimination for lifted weights is abnor- mally lowered. To carry the analysis further, how- ever, is not advisable until the scale has been more accurately ^'calibrated." When this has been done, its value for individual mental diagnosis will be greatly enhanced. While, therefore, the peculiarities in our curve can be partly ascribed to extraneous circumstances and to imperfections in the scale, they also implicate in- herent anomalies in the mentation of the epileptic. In fact, our B.-S. testing has furnished us with a picture of the results of a process of mental wreck- age caused by pathological processes going on in the afflicted individual or by a neuropathic heredity, whereby the integrity of various mental functions has been impaired at various levels of mental devel- opment, or whereby certain lower levels of mental TESTING THE BINET-SIMON TESTS 55 functioning have been swept away, while higher levels have remained intact — ^mental sentinels that have remained to tell the story of destruction. This interesting clinical picture of mental ruin and havoc is still incomplete in its finer delineations and invites further psycho-clinical research. (4) Finally, the facts brought to view above should make it patent that the B.-S, scale is still in its experimental stages — a fact that should occasion no surprise when we consider the short time that it has been in use, particularly in this country. It still requires a thoroughgoing tryout and revision. Ee- visions have, indeed, already been attempted by Binet and Goddard. In some respects their changes harmonize with the requirements of our comparative table (Table VII), but in other respects not. In this table only the tests most obviously misplaced for epileptics have been compared with normal perform- ances.) Since there is considerable disagreement in the conclusions of the five writers who have worked with normals in France, Belgium, Germany, Eng- land and the United States, it is obvious that the question cannot be considered closed. American in- vestigators will have to establish norms for Ameri- can children. The studies already made sufficiently indicate that a test too difficult for children of one nationality may be too easy for those of another. For American workers Goddard 's results, based as they are upon an extensive testing of American chil- dren, are the most suggestive. But it appears to me wise to continue the use of the 1908 scale until more returns are in — until we have tested large masses of boys and girls by the ivide-range testing to which I have already alluded. It is questionable whether we 56 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES can calibrate and improve the scale by confining the testing to a very limited range of years. [Obviously, this work will require the earnest cooperation of many psycho-clinicists. More than that, it ought to receive the undivided time and attention of child study specialists — of a worker or group of workers who can devote themselves to this particular service as a life career. In my judgment, this work can probably not be done effectively and within a reasonable time unless it is established as an independent department of research. The clerical labor connected with investi- gations such as this is so onerous that it should be cared for by a force of clerical assistants. At the present time all, except possibly one, of the psycho-clinical investigators are swamped with purely clerical work (computing, tabulating, etc.), and must use up their best energies in purely routine and mechanical labor. This is uneconomical from the standpoint of public finance and the conservation of the nation's best creative brain power.] The further improvement of the scale, we may add, requires the transposition of certain tests, the elimination of certain others which seem valueless, the equalization of the number in each age or the proper differential adjustment of the system of ac- crediting points, and the increase rather than the decrease of the number of tests for each age. I feel that the last point requires emphasis. We need to probe a considerable number of traits and capacities for the reason given (p. 42) if we would arrive at a true clinical picture of the child or a correct classifi- cation. We cannot hope to diagnose or grade accu- rately if we test only a few capacities, or if we base our judgments on only a few symptoms. The ele- ment of variation in human traits is too large. What is needed is a well-balanced, comprehensive survey of the fundamental human mental capacities. The minimal number of tests for each age should be placed, I believe, at ten. I should regard it as unfor- tunate to eliminate such tests as the age, writing and reading tests. All of these have furnished important information with respect to the epileptic (as well as TESTING THE BINETSIMON TESTS 57 various types of insane patients recently tested). Nor is it essential to eliminate all the tests which are dependent upon training (or, indeed, all which are pedagogical), partly because tliis is not desirable and partly because this is impossible. Nature and nurture proceed hand in hand, inseparable, recipro- cal, interacting and independent only in conception. Just as we posit a normal rate of development which the forces of human nature undergo — normal, that is, within limits — so we may posit a normal curve of development within variable limits for a given order of civilization or social evolution, which human changes follow as a result of the processes of nur- ture. We cannot, if we would, test merely pure na- tive capacity uninfluenced by environmental agen- cies, except possibly during the first months of life. But we can measure native capacity as modified by the environment. In this country, where we have fairly uniform standards of educational require- ments in the schools, there would seem little rea- son for eliminating some of the more funda- mental academic tests. A child of eight or nine who cannot read at all certainly demonstrates his intel- lectual inferiority thereby. To repeat : The evidence regarding the imperfec- tions in the Binet-Simon measuring scale of intelli- gence cannot be brushed aside. The scale certainly has not yet been made maximally accurate or scien- tifically precise — scientifically precise in the Aris- totelian sense (we may demand only that degree of accuracy in a given subject of inquiry which the sub- ject-matter itself allows) or sufficiently standardized. A large work and a challenge remain for the students of "Intelligenzpriifung" — a work that will pay 58 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES double interest on a large investment of time and labor, for the public schools, juvenile courts and in- stitutions for defectives stand in dire need of a sim- ple, objective, practical mental measuring rod by which to determine the degree of mental arrest, de- fect or acceleration of deviating individuals, and by which properly to classify institutional cases. In the light of the defects which we have found in the scale, what, then, shall we say of the value of the B.-S. system of classifying defectives? An answer to this question must await the analysis of the fol- lowing chapter. CHAPTEE III. THE VARIATION OF MENTAL AND PHYSICAL TRAITS IN RELATION TO THE AGE CLASSIFICATION OF THE BINET-SIMON SCALE. Paidologists have been wont to accept tlie postu- late that mental traits or capacities increase in effi- ciency or multiply in number with increasing age. The validity of the B.-S. scale itself depends on the validity of this assumption. If we accept the as- sumption as correct, the following conclusions follow : First, it is feasible to construct a graded scale of mental performances consisting of a series of tests which either progressively increase in difficulty from year to year (similar tests being repeated at various levels) or measure new traits developing at various higher levels. Second, it should be possible to arrange these tests in a fairly accurate ascending age series, so that we can locate the mental station of normal and ab- normal individuals in units of mental age, and deter- mine how any mental or physical characteristics vary in accordance with a fixed classificatory or graded scheme. Third, by plotting age-curves for the individual traits tested in the scale (as well as for functions tested independently of the scale), we secure a means 59 60 EXPERIMENTAL STUDIES OP MENTAL DEFECTIVES of trying out the accuracy of the scale itself and of determining more fully to what extent these traits vary with age among normal and abnormal indi- viduals. In this chapter I shall be concerned with this third conclusion, because the imiDcrfections revealed in the 1908 B.-S. scale by the data considered in Chapter II naturally raise the question whether or not the scale is of any practical value in the mental grading or diag- nosis of normal or abnormal subjects. If it can he shown that the average efficiency in the various traits tested increases ivith growing age {annually or biennially), then the scale, in spite of its imperfec- tions, and provided the mean variations in each age are not too large, remains a valuable means of deter- mining, in comparable terms, the classification of different homogeneous groups of persons. For if the individuals were very poorly classified, we should expect to find obvious irregularities or skews in the curve of efficiencies for each trait, and extreme mean variations. Of course, the relevancy of the scale, even if it be imperfect, for determining the relative station of different individuals in the same homoge- neous group, needs no argument. The discussion of the scale will also serve to bring to view various facts of considerable interest re- specting the mental efficiencies of epileptics. In order satisfactorily to analyze our data it will be necessary to examine the averages for the entire population, of all ages and both sexes, in the various tables, and the corresponding mean vari- ations. The former will be examined with consider- able thoroughness, the latter more briefly. Finally, VARIATION OP MENTAL AND PHYSICAL TRAITS 61 a concliiding section will be devoted to the more general conclusions and comparisons. It is unfortunate that, in the absence of the origi- nal data, it will be necessary to limit this study to four B.-S. tests, and to tests with the form-board, dynamometer and ataxiagraph. Explanation of Tests. In the color test (Table X) the patients (224 epi- leptics between the B.-S. ages III and XIII) were uniformly, with few exceptions, instructed to name the four colors as rapidly as possible. The experi- menter pointed to each color in succession rapidly, so that the results would represent maximal per- formances were it not for the fact that the green used in the testing of about one-half of the patients was poorly saturated, and therefore caused hesitation or failures. Green, even at its best, is the least fre- quently named of these colors, among normal per- sons, as appears in Bobertag's B.-S. testing of Ger- man children, while red is the most frequently named. Blue and yellow were about equally difficult. The following was found by Winch to be the order of the correct application of color names among children three to five years of age: black, white, red, blue, green, yellow, violet and orange. The order agrees with Bobertag's, except that green is placed ahead of yellow. In my testing of epileptics no other single color caused so many delays, indecisions or failures as green. I agree fully with Bobertag that these colors, in order to serve a useful purpose in the scale, should be standardized. In the test of uttering discrete words the subjects were always told to utter just as many single words 62 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES as possible during three minutes or until told to stop (Table XV). Those who were slow were usually incited to greater effort by verbal suggestions. Three or four examples, illustrative of the object of the test, were furnished at the outset. No record was included which consisted merely of phrases or sen- tences or which consisted of any considerable num- ber of these, and none was excluded because of pecu- liarities in the word lists (such as uttering many numbers — unless, indeed, all the words were num- bers — or letters of the alphabet, or names of months or days). In the reading test (Tables XI and XVII) the fol- lowing selection was used: Three | houses | on | fire. New York, | September 5th. | A fire ' last night | destroyed | (three houses) in the center of the city. | Seventeen | families | are without homes. | The loss is more than ] thirty thousand dollars. | In saving | a child | who was asleep in bed, | one of the firemen | was badly | burned | on the hands and arms. | (The units are indicated by the line divisions. In a few cases half credits were given ; e. g., half a memory for "September," "on the hands," "thirty thousand.") The subjects were merely told to read the selec- tion, without intimation that they would be expected to reproduce it. To many this test was given after some of the speed tests, so that a number of subjects undoubtedly inferred that they were expected to read the passage as rapidly as possible. The results may therefore be a trifle better than the general average. In the form-board (Table XIII) and hand dyna- mometer (Table XIX) tests the patients were, again, invariably urged to do their best, and only their best records are averaged in the tables. Each one was given three trials each with the Vineland pattern VARIATION OP MENTAL AND PHYSICAL TRAITS 63 form-board and the Smedley dynamometer, but the tests were repeated during second or third sittings for many subjects, so that frequently the figures utilized are the best single records in from six to nine trials. It may be said, parenthetically, that the best form-board record was by no means always the last one ; on the contrary, many epileptics lost with prac- tice or familiarity. Uniformly, the form-board was placed directly in front of the subject, with the side containing the star nearest. The subjects saw the blocks removed from their proper places by the experimenter. The blocks were thrown into a heap on the table along the far- ther side of the board, instead of along either end, so that they should be equally distant from the two hands. Either hand could thus be used with the same ease. The subjects were told to use only one hand. The same uniformity of procedure was used in the dynamometer test. The instrument, properly ad- justed to the size of the hands, was placed by the experimenter, dial up, in the subjects' hands so that the palms always faced up. Two results thereby fol- low: (1) the subjects were able to see the movement of the registration pointer; and (2) the thumb sides of the two hands, which probably exert a greater pressure than the little finger sides, came into con- tact with the opposite ends of the stirrup of the dyna- mometer. This circumstance may be of some impor- tance in dynamometry work, as indicated by a series of calibration tests of the instrument, which showed that the same weight on the two sides of the stirrup does not give exactly the same reading. The weights were suspended from the stirrup by an iron hook, so made as to move freely (without friction against any stationary part) Objective Weight Result of Normal Lift. Kg. Right. Middle. Left. Rigt 5.2 4.7 5.2 4.7 12.4 10.3 11.7 10.0 16. 16.5 14.1 16.4 14.3 15. 31.7 28.2 30.6 27.4 48.2 45. 45.7 35.5 60.3 48.8 58.2 64 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES when the dynamometer was lifted from the dial end. Thus pressure was exerted on the stirrup in much the same way as when it is squeezed by the hand. The hook was suspended from one of three points on the stirrup : the middle and either end. "Normal" indicates that the weights were lifted slowly or with moderate rapidity ; "fast," rapidly or with a quick jerk. (The weights were weighed on a new pair of platform scales, so that the weighing is probably correct.) TABLE IX. Calibration Tests of the Smedley Dynamometer. Result of Fast Lift. It. Middle. Left. 19. 24.7 16.1 36. 64.9 The figures are averages of from three to twenty trials. A number of interesting fact^ appear from a study of these figures. (1) Pulls on the right and left side of the stirrup practically always gave a different reading, usually amounting on the average to a frac- tion of a kilo (but with a maximum of nearly 10 kg.). In most in- stances the right side of the stirrup gave the larger registration. (2) The middle pull in all cases gave the highest registration — presumably because the pulls at the ends occasion friction along the sides of the piston. The difference between the middle and side pulls ranges be- tween 1.4 and 3.2 kg. (two extremes excepted) — a difference so large as seriously to threaten to vitiate dynamometry work. Consequently it is of the utmost importance in testing with the Smedley dynamom- eter that the subjects exert the pressure along the middle of the stirrup. This precaution is by no means always observed, particu- larly not when readings are taken while the handle has been allowed to slide out of the palm on the side opposite the thumb. In this case the pressure is exerted at the end of the stirrup, with a consequent less. In the present research the attempt was made to secure middle pressure. This instrumental defect should be remedied. (3) If we consider the middle registrations, it appears that the dynamometer in use was not true to scale, except for one or two weights. The spring was slighty too stiff. The inaccuracy, moreover, seems to in- crease with increasing pressure, as shown by these differences for the successive weights (beginning with 12.4 kg.) : .7, .1, 1.1, 2.5 and 2.1 kg. In the interest of scientific precision, dynamometers must be calibrated as accurately as possible and tested objectively from time to time. (4) The testing with the same weights sometimes differed quite considerably on different days (the tests were made be- tween December and March). This may possibly be due to tempera- ture or other atmospheric changes which may alter the tension of the spring. Is this a remediable defect? (5) The differences between the VARIATION OF MENTAL AND PHYSICAL TRAITS DO "normal" and fast or jerky lifts were large, amounting, for the "mid- dle" readings, to 7.3, 8.3, 5.4 and 6.7 kg. for the different standards. (The friction of the recording pointer was kept uniform during all the tests.) The quick lifts or jerks, of course, invariably gave higher readings. It is therefore evident that dynamometry results may he invalidated, for purposes of comparison, hy differences in the method of squeezing. I have found that most subjects tend to apply a steady, gradual pressure. But there are many who give a rapid or violent squeeze. Nearly all of our tabulated results were obtained by the first method. The readings, however, according to the above figures, appear to be somewhat too low. It should be added that no results are tabulated from subjects having any but slight injuries in one or both hands or arms. Finally, in the ataxiagraphic test (Table XXII) the subjects were required to stand under a small horizontal sliding board, to which was fastened a tracing record-sheet of white paper. The board could be moved vertically, so that it could be adjusted to any height of subject, and also sagittally, so that the ataxiagram could be properly placed on the paper with reference to the antero-posterior axis. The ataxiagram was made by a soft marking pencil, which was glued to a band of spring metal, attached to a sort of ataxiagraphic helmet, which consisted of two curved encased metal bands fitting snugly over the cranium. The spring kept the tracing pencil in con- stant contact with the paper, except when the subject dropped the head or swayed too much. In such cases the experimenter was sometimes able to lower the tracing sheet, or if not, to trace the record by hand with sufficient accuracy. In a number of cases in which the ataxiagram extended beyond the paper it was possible for the experimenter to determine the extent of the departure. At the beginning the spring was depressed by the experimenter's hand until the record board had been properly adjusted. The start- ing point on the ataxiagram, the point of initial con- tact, or a point not far removed from this point, 66 EXPERIMENTAL STUDIES OP MENTAL DEFECTIVES would have afforded four measurements, anterior, posterior, to right, to left ; but the data for this tabu- lation are not now available, so that in Table XXII only the gross measurements are given for the an- tero-posterior and lateral sways, irrespective of the point of initial contact. Two methods were employed in this test. In the first, the subjects were instructed to stand with the feet placed to form a V (heels together). In one series the eyes were closed, and in the other the open eyes were allowed to rove at will. In the second method they were instructed to stand with both heels and toes close together, with hands hanging along the sides, and the eyes, when open, fixated on a black disc (about 1 inch in diameter) attached to the win- dow curtain about 12 feet away. The disc was ad- justed to a level with the subject's eyes. In neither method were the subjects instructed to try to stand still or rigidly, but simply to stand naturally. Noth- ing was said regarding the object of the test. With both methods about half the subjects were first tested with the eyes closed and about half with the eyes open, and each test (given in close succession) lasted exactly one minute. Table XXII contains the data secured by the second method only. A search of the available literature indicated that there has been no uniformly observed method in this test, but the sec- ond method here described seems to the author to possess most merits. The limiting of the tabulation to the second method explains, in part, why no more records are used in the study (104 males and 40 females) : the station of many patients was taken by the first method. Another reason is that ataxia- grams secured from ataxic, choreic, paralytic, para- VARIATION OF MENTAL AND PHYSICAL TRAITS 67 plegic or hemiplegic subjects were excluded from the tabulation. Accordingly, the ataxiagrams used should represent the body sway in epilepsy uncompli- cated by special permanent motor disturbances. It should be added that a number of patients, particu- larly low-grade ones, either did not fixate the disc at all, or allowed the eyes to rove more or less. Analysis of the Avekages. In Tables X, XI and XIII an increase of efficiency or capacity with increasing B-S. age is shown by a progressive diminution of the time of execution. In the corresponding graphs (II, IV and VI) this is shown by a drop in the curves. The Time Needed to Name the Four Colors — Red, Yellow, Green and Blue. The time required to name the four colors. Table X, does, indeed, decrease with increasing B.-S. age, but the decrease from year to year is not very regu- lar, as seen at a glance in Graph II. There are numerous exceptions in the averages for the general population, the children and the adults. The excep- tions are least numerous among the girls and women. The differences, however, between the groups, the imbeciles and morons, for all patients, 4 seconds, and especially between Age III and Age XIII, about 7 seconds, are quite considerable. The difference be- tween the averages of Ages VI and VII and of the moron group is much greater for the adults than for the children (3.7 sec. compared with 1.1); for the girls than for the boys (.9 as against .2), and for the men than for the women (6.8 as against 3.1). The sex differences are also brought out by the general C3 68 lo = 00 1^ *! •> « K «» 1-4 0> to 00 c^ >n 00 la o> CM r^ -< .-« N la 00 00 to cj '<«• ci e>» 00 o> •-; «» «> CO i-i e^ 94 C4 «^i e4 ^ ^ eo CO t^ f^ •>■ Cq CO <=> 00 f^ V to t>^ « f-i u> i-l .H 11 CO O * ^ = > > > ^ » 5 X X •< > "" ^ S I S 4> X < I 0) ■6*' «5 V 9 £3 t^ 3 Ja Li ^ ©a; wo b5 "a >• 4) ag: , a y tj o 4> n o S o « ? HjaDO _ »- >5 73 -« B ■3 §-«?* .ogfca 0-3 ® 3 '-' M 91 O) «S S 60 3 0) "-- , 000 -tJ-J !; a y o r, o— " . UiS * >• OB'S •1 t- flB^ •5mc .w-oog-ssa •"'^•S a^„— *•— 73 ID a cSi-c—tH^ ^ •OS O .. <^, g * e« 3 ^ 2iS 3 a« rt «i2 fee * > '^ B! 0) tf-: O rt-M « gS'So-aajgag tig. 54,*«53 a) -tJ S m"© fe *-i c^ -o j3 aS rt S . 3 3 S © > M a btiTS q -a) y «, *^* * ^: » a tHg:ii :?'<;! J-S S-3 «fl VARIATION OF MENTAL AND PHYSICAL TRAITS 69 averages for Ages VI to XIII, from which it appears that the girls are superior to the boys (average of 4.9 see. compared with 5.7) ; the women to the men (5.2 compared with 7.9) ; and the children to the adults (5.7 compared with 6.5). In such a simple trait as the time of naming four colors it may be as- sumed in harmony with the above findings that in- creasing maturity will not accelerate the speed after the colors have once been really learned. The following conclusions seem to be justified : (1) Significant sex and maturity differences (dif- ferences between the juvenile and adult periods of life) are brought out in so simple a test as the speed of naming the four fundamental colors. Epileptic children are superior to adults, and girls and women to boys and men. That normal girls excel normal boys in the knowledge of colors has been shown be- fore. Bobertag recently found this condition to ob- tain in this very test. Moreover, it is stated that color-blindness is more prevalent among males than females (in about the proportion of 4:% to .5%). There is a bigger difference between high and low grade adults than between high and low grade children. (2) This trait (the speed of naming the colors) apparently reaches its maturity at about Age IX. (3) In this test the increase with each increasing B.-S. age is not very regular, indicating either that the subjects are not very closely classified or that this is not a very satisfactory test by which to check the accuracy of the scale. eg- 15 o 6 > is t-l . 2a LZ in iH t- 00 o> N lA o:) ^ lA e-; N lO la th to ■* c4 CO us CO M C4 o od t-> lO «o ^ M »H iH iH to lO tH ?5 N I ■* •♦ t- ^ 00 U5 O O M t- CO lA »H 00 ■^ CO M lA o o o ■^ t-^ to CO OO CO CI cq ICO c^ ■* t- 00 lA ■^ N Tf o to ■* rH r-l M r-l © M 0> lA O to lA iH ^ CO CO iH O lA cq N Oi rH lO t^ CO to CO ©» 1-; t- CO Til lA ia' to rJ ■* 00 to CO M CO to CO ci lA CO 1-i <»• 3 T-l >-* «> © lA O CO © ci C5 M ^ CO t-l © © © to in ■* t-^ 00 ^ T}" -^ tH •* CO to CO OS W CI 00 © E-^ Otj F-j CO lA lO >-l © © © lA lA eo' »H lO lA 00 CO CO I M to lA 00 CO 00 cq 3 o to s 51 00 «. 1 •H CO j:; s; s S S M >H >H 61 >- (Q-< « a lO 00 s ^ a ^ 1 1 a > i-i '^ < § «« ?> a) BO ^ O) 0) S i. OS a>*» •s-=i u u ^ +j o *3 — > a* * ""O bg . *y VARIATION OF MENTAL AND PHYSICAL TRAITS 71 The Time Required to Read the Passage About a Fire. On the other hand, the time required to read the selection, Table XI, decreased appreciably and regu- larly (as is strikingly apparent in Graph IV), with only one exception among the girls, and two each among the men and women. The reading time ranges from 129.2 sec. Age VII, to 18.5 sec. Age XIII, a difference of 80.7 sec The difference be- tween the averages of Ages VI and VII and of the moron group amounts to as much as 56.1 sec for the general population, 111.7 for the boys, 50.8 for the women, and 30.1 for the men. The shortening of the time from each age from IX to XIII amounts, re- spectively, to 24.6, 17.3, 18.1, 3.3. and 4.7 sec— a series of progressive improvements. The largest and most consistent improvements in these years are made by the boys and men (ignoring the girls and children, where the data are incomplete), as seen in Table XII. TABLE XII. Improvement in Reading Capacity with each B.-8. Age. (The figures represent the gain in seconds made hy a given age over the next preceding age, based on Table XI.) Age. Boys. Girls. Children. Men. Women. Adults. IX 49.6 81.0 77.5 31.1 .4 14.0 X 33.1 — 3.0 26.3 16.8 4.6 15.6 XI 35.7 28.4 8.0 13.7 14.6 11.9 XII 11.8 8.7 —12.6 1.7 XIII 7.5 — 1.0 12. 4.0 Ave. 27.5 35.5 37.2 13.8 3.8 9.4 A minus sign indicates loss of eflSciency. From Table XI it is apparent that, in respect to reading ability as gauged by time, the girls are di§- 73 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES tinctly superior to the boys ( shorter time in all ages and in the averages) ; the adults to the children (in all ages and the averages), and the women to the men (in six of eight ages and in two averages). We may accordingly conclude that the reading test is a valuable test for four reasons : (1) It discloses significant sex and maturity dif- ferences. Here the epileptic females surpass the males and the adults the children. But the difference between the high and the low grade groups is great- est for the boys. (2) The considerable superiority of the adults indicates that reading is an adult art, and that it seems to be worth while to teach reading to epileptics in spite of their tendency toward progressive deteri- oration or degeneration. This point, however, re- quires special investigation. (3) Apparently there is a large age-difference in reading capacity, particularly from VII or VIII to XI. The very considerable gain at IX seems to in- dicate that there is a pronounced advance in reading ability for epileptics at this age. At the same time it is seen that the progressive decrease in the reading time is fairly regular from VIII to XI. (4) A reading test thus supplies a valid test for differentiating mental capacity and for checking the accuracy of intellectual measuring scales — albeit the test is more pedagogical than psychological. As judged by the time of reading, the B.-S. grouping appears to be fairly satisfactory. It seems, therefore, unwise to eliminate the test VARIATION OP MENTAL AND PHYSICAL TRAITS 73 from the scale, as Binet has done in the latest (1911) revision. So important was the reading test in the original B.-S. scale that it was used as the differential between two groups of defectives, the imbeciles and the morons — the latter of whom can be taught to read, while the former cannot. The fact that the test is more pedagogical than psychological, and that it represents the results of training, should not neces- sarily militate against its use for psychological diag- nosis. The attempt to teach children to read will probably have been made in the case of the majority of subjects who will ever be tested. The Time Required to Replace the Blocks in the Form-Board. In this test (Table XIII) there are decided and consistent gains (as is markedly apparent for the general population in Graph VI), with three excep- tions each among the boys and girls, and two each among the men and women. The difference between Ages I and XIII amounts to 211.6 sec, and between the imbecile and moron groups 84.4 sec. The differ- ence between the imbeciles and morons is consist- ently larger for the girls than for the boys (97.7 as against 78.7 sec.) ; for the men than for the women (63.6 against 50.9), and for the children than for the adults (89.1 as against 63). On the other hand, the male morons are five times as efficient as the male imbeciles ; the female morons three times as efficient as the female imbeciles ; the boy morons 5.5 times as efficient as the boy imbeciles, and the girl morons 5.2 times as efficient as the girl imbeciles. The corre- (l V to m ^ o ▼- fH C4 «-l a^ '^ "i-i. B5 w 00 0) >> •OS •-' «> 1 «- « rH ^ [3 a ^ ■a ■ o ^ -^ji o6 M -10 0) 00 CO ri O --i ■* CO 00 OS •» lO <0 1-1 e^ c> ■«; CO o> 00 ^ ci ci t^ *^ CO ^ flj ri ,< ,i< rH »- r-l m Oa ♦-< iO CO "^ *T ---2 > 9 <1® W CO C4 u^ ^ 0> c^ ^ ^ r-i *- <"> in CO eo *H 5& 5: CO O C^ M »-» 1-1 ^ ^ iO A V 00 CO «0*000C0fOd0J copmcof^'Wt-^gj ■♦NC0C^rHiHF-1« «0 lA O A C0 CD ^' ^ ^ C4 e» CO iH •<«• S TJ C > J) Oi » S 3 2 t» m +j 13 i> h-c if a »><* O) p (* O 05 o >> fl «i J' M-S 3 0)="^ M > a, , °*" .-I -OCB-MaS^'O'" si 9!'"ai^-«l^„, ti £ a ajif a '^ t-T3 Mi« §" •? * S25 a Sa V*J > © 5S t^ 1 ^:f: SQ a ~ ° > 15 -o& ■H M »- L^ s .I M T- 5 & 00 a, a p ed p EO 0-3.2 C t, «H © 4( " "->» a) aj <" ■!-> -M S «* es w dl ..« 0) > K > Ml rt 0) a; 08 WO >- i; t- 0) > 1) bo's q; aj t, « 3 > «* ^ si w a '- t-5 i5E o. P ts 4, aj © _+J (4 »- t. S « * 33 ^n-i "as. £ a a 41 o §! S ^ < > VARIATION OF MENTAL AND PHYSICAL TILVITS 79 lation : 11.9, 10.2, 7.7, 8.6 and 5.1. The correspond- ing figures for the separate groups are shown in Table XVI. TABLE XVI. Gains or Losses with each B.-S. Age in the Numher of Words uttered in Three Minutes. (The figures indicate the difference in the num- ber of words given in a given age and the next previous age, as computed from Table XV.) Age. Boys. Girls. Children. Men. Women. Males. Females. Adult IX 2.7 17.2 8.7 15.4 12.7 10.8 14.1 13.8 X 10.7 18.5 13.2 11.1 5.9 10.9 9.0 8.1 XI 9.2 5.9 10.2 2.4 9.0 5.1 12.3 6.7 XII 13.0 6.0 15.3 7.3 3.0 17.4 7.3 6.0 XIII 4.5 8.8 11.6 —4.4 1.0 9.3 Ave. 8.0 11.9 11.8 9. 8.4 7.9 8.9 8.7 Here the gains are the highest for the girls and children. They continue, though in diminished de- gree, to Age XIIL This indicates, as one might ex- pect in a test of this sort, that the rate of utterance increases beyond the age of XIII, so that a free asso- ciation test might be found valuable at higher levels. (Possibly a controlled association test would be still more valuable at various levels.) In the rate of word utterance the girls excel the boys (better in five of the ages and in the three averages) ; the men the women (better in seven ages and in two averages) ; the children the adults (better in five of eight ages and in three averages), and the males the females (in six of eight ages and in two averages). The following conclusions follow : (1) There are certain sex and maturity differ- ences in the free association test, although the sex- differences do not appear very clearly among these epileptics. (We are speaking of quantitative, not qualitative differences.) The girls seem to surpass the boys, the men the women, and the children the > lA CO t^ rH CO eo rH CO ,, ©2 r-J jA r-i rH r-i tA e^ r-i |1^ iH ^T 00 CO lA 00 ^ ■»»< N ci eo TjJ l« d d iri t-i in -o •- . CU d CO iM t- 00 rH t- . 55 iH ■«< cq t- cq rH w-t :« &M r^ t«; lA «5 t-; CO M t~ eo Is HI rH r-i i-i r-i rH evi r-i M lA CO lA lA lA 00 d a 0) >-^ 1H .0 1 * C5 Oi t- M" to lA tc p "« [^ r-( rH CO 05 .""■S-o >-a2 '> cc ?J 00 CO CO ■^. to esi .0. ►"H ^ d r-i T-i r-i j-i r-i r-i f-H r-i • Ma; •--1 •^ S" S lA Irt 00 00 to K. t- ^ .-2 s T-i e<5 lA lA d t-i 10 t-^ in a 0) 3 d •^ M CO f.H rH t- f-t 00 flo.2 ^!; W rH CO C4 rH r^ ©"M m irt rH lA r-i «o t-; r-i 00 ■* +2 M ,^ eS 2 ^ 5§; 1-H lA OJ ^ CO r-; rH 9> 1' L< N eo ■^' ■<«« d lA lO t-i 1= a ajHH 0) 0)'^ '§ r^- 00 CO rH rA '-' "-I r-i •"* C^ r-i 0^ 0. ' ^"^ 7« ■»»< lA lA 00 t> K 00 .m ^ 10 CO kA 00 «9" >A lA CO '^ d t- ^ Sm-3 2-^ '^ rH CO -r^-o 0) eS I" '> (M T-H Fl' rH r-i f^ jA -oca lA 00 d rH d q d rH <0 in in .9 fl r? OJ u ^ S > o> t- 00 00 pa rH q Vj "♦^ S d d r-i r-i d r-i r-i (1 +3 . 4> i. '^s; w t- «D CO t-; «o w OS m^M .. td oj ej w •fl< •>«< ■^ t-i * 11= s 6 'S" ^ CO 00 eo rH ?% $ -ft** /> tA «D CO kA M *i q h^ rH ^ e^ r-i r-i rH «: cal* 17 -t-) .^ ra >--^ 10 C5 r~; to lA o> t- in rH CO ■>»• 06 d ^ > X X >< X 1 >< > VARIATION OF MENTAL AND PHYSICAL TRAITS 81 adults. The gains with advancing years apparently are greatest for the girls and children, while the dif- ference between the high and the low grade groups is greatest for the males and the children. (2) There is a fairly regular, although slightly diminishing, increase in the capacity with each rising B.-S. age. (3) Accordingly, the test is of value for mental classification and for checldng the accuracy of intel- ligence scales. (4) The B.-S. scale fares fairly well under the scrutiny of this test. Facts Reproduced in the Reading Test. The number of facts retained in the reading test likewise increases regularly (Table XVII, Graph V), with one exception each among the boys, girls, women and children ; two among the men and adults, and none for the general population. The increase ranges from 2,1 at Age VII to 7.4 at XIII, a differ- ence of 3.1 memories. The gains for each successive age from VIII to XIII are, for the entire population : 1.6, 1.1, 0.5, 1.2, 0.3 and 0.6. The figures in the sep- arate columns are as follows : TABLE XVIII. Gains or Losses with each B.-8. Age in the Number of Units Repro- duced in the Reading Test (the figures indicate the difference be- tween a given age and the next previous age, based on Table XVII.) B.-S. Age. VIII Boys. 2.4 Girls. .4 Children. Men. 1.5 1.9 Women. 1.1 Adults. 1.4 Males. 1 2.0 femal( 1.0 IX .8 1.9 1.2 1.9 1.0 1.4 1.5 1.2 X 3.3 — .3 2.1 —1.0 .3 .3 .8 .1 XI —1.4 .4 — .7 2.6 .9 2.0 1.0 .8 XII 4.9 2.3 3.9 — .8 — .7 .9 .8 .1 XIII —1.8 1.2 1.8 1.4 .1 1.3 Ave. 1.3 .94 1.8 .96 .7 1.2 1.0 .75 82 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES The increase in retentiveness thus varies consid- erably from age to age and group to group. The largest average improvements from VIII to XIII are made by the boys, adults and the males. This coincides with the highest memory efficiency, as de- termined by the number of units retained : the boys surpass the girls (in five of six ages and in the aver- ages) ; the children the adults (only in two of five ages, but the averages are higher) ; the men the women (in four of seven ages and in the averages), and the males the females (in six of seven ages and the averages). These facts suggest the following conclusions: (1) The ability to reproduce prose (read by the subject) varies with the sex and the period of life. Here the males and the adults excel in the average amount of improvement from year to year, and the males and the children in the absolute number of units reproduced. (2) There is a diminishing, though not entirely regular, increase with each increasing B.-S. age. (3) The irregularities found in the age curve in the various columns are due either to the fewness of the subjects, to the peculiarities of the groups (epi- leptics), or to defectiveness in the B.-S. classifica- tion. In general, however, the results speak more strongly for the validity of the B.-S. scale than against it. (4) The test, in any case, is a valuable test for purposes of diagnosis or classification. Whether it is preferable to base the number of facts reproduced on the logical units, following B.-S., or on phrase or thought unities, has not yet been determined. The scoring can be made both qualitative and quantita- VARIATION OF MENTAL AND PHYSICAL TRAITS 83 tive. Possibly the best plan is to base tbe scores on single-idea unities, and scale the scores according to the relative significance of the items (a combined qualitative and quantitative measure). Mamiometry. The strength of grip (Table XIX) likewise in- creases with the degree of intelligence, although the gain with these patients is by no means always regu- lar, as is most easily seen by referring to Graph VII. For the right hand there are four exceptions each for the general population, boys and girls ; three each for the men and adults, and two each for the children and women. For the left hand there are four exceptions among the women and girls, three among the men and adults, two among the boys and entire popula- tion, and one among the children. Apparently the gains are more consistent for the left than for the right hand, and for the children than for the adults. By tabulating, as in Table XX, the amount of the gains and losses in each age from V to XIII, we not only see more readily the irregularity of the increase from age to age, but we also perceive that the sums of the average gains for the boys, girls, men and women, after the losses have been deducted, are somewhat larger for the right than for the left hand (8.4 as against 5.5 kg., whence the more intelligent, the stronger relatively is the right hand) ; that the children gain more relatively with the right hand than do the adults ; that the boys gain relatively more with the left, but the girls with the right hand ; that the men gain more than the women, and that in all cases the strength of both hands is less for the XIII- year-olds than for the Xll-year-olds. *- CO 00 ?* Q CM C4 CM CM CM CM M ^ SQ „ r- Tt !S ? SQ CM — ,< > > = J = K X r: VARIATION OP MENTAL AND PHYSICAL TRAITS 85 . t-^ lO CO -^ CO iH us 'id "O OOrHOqOOC-OOOl-IC- <] " " «» te r .oiusirt-^coc^i^oc^io rSfi r« 1=^ 1 1 1 &-•" L 1 1 1 wis s . e f "^ '"' <^ "? '^ "^ "^ R ^ 1 J- CO -* ■ ' -* rH rH «J l> 1 INI 55 g S. usoooocoe^t-McoLft J *COC I ftj*^ oq !» ■^ CO ' r-i CO tH s S-o U Di ■•J' iH us " -*' -^' [-■ CD 00 ■* cx) CO c<; us CO CO t-H CO T^ us ci O *~ ^ J3 ,COTt< <1 86 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES It would be important to know how far these pecu- liarities or irregularities are due to the small number of subjects tested in certain ages. The difference between Ages III and XIII for the entire population is 12.9 for the right hand and 15.3 kg. for the left hand; the corresponding differences between the imbecile and moron groups are 11.6 and 12.2. That the gain of the morons over the imbeciles is not uniformly greater for the left hand is seen from the following tabulation : TABLE XXI. The Amounts in Kg. hy which Morons excel Imhemles in Strength of Grip (based on Table XIX). Right Hand. Left Hand. Boys 13.1 kg. 14.1 Girls 10.6 7.9 Children 13.6 13.2 Men 9.3 11.2 Women 8.0 5.5 Adults 10.8 10.2 The gain of the high-grade over the low-grade patients is relatively larger for the boys than for the girls, and for the children than for the adults. With respect to the comparative strength of the right and the left hands, the evidence is conflicting. The general average for the left hand is .5 kg. higher (index of right-handedness = 1.01%). It is higher for the boys (index = 1.03%) and women (index = 1.02%), and lower for the girls (96%) and men (98%). Among the morons it is higher for the boys (1.02%) and men (1.01%), and lower for the girls (92%) and women (99%) ; and among the imbeciles higher for the girls (1.03%) and women (1.10%), and lower for the boys (99%) and men (96%). The index is about the same in the two groups. The aver- VARIATION OF MENTAL AND PHYSICAL TRAITS 87 ages for the boys are higher for the left hand in five individual ages, lower in three, and equal in three : for the girls they are higher in half the ages ; for the men, higher in six ages and lower in five; for the women, higher in seven and lower in four; for the children, higher in half the ages, and for the adults they are higher in six ages and lower in four. It is apparent, therefore, that the left-hand grip is stronger in some epileptics and the right-hand grip in others, with the odds rather in favor of the left hand. It has been claimed that degenerates have a stronger left-hand grip, but the rule is by no means invariable, at least so far as epileptics are concerned. For normals the right-hand grip has been found superior, the index varying from 81 to 96%. This contrasts with our index for the entire population, 1.01%. But there are probably exceptions among normal persons also. If the dynamometry results leave us in doubt re- garding some points, the superior strength of the epileptic boys and men is unmistakable, just as among normal persons. There is no significant ex- ception in any column. The superiority of the boys, compared with the girls, amounts to 8 kg. for the right hand and 9.7 for the left, and the superiority of the men, compared with the women, amounts to 16 and 14.9 kg. for the two hands, respectively. The difference is not only large, but the male superiority is relatively greater among the men than among the boys. Among normals, likewise, it has been found that the divergence grows more marked from the time of puberty. Likewise the difference between the moron males and females is greater than between the imbecile 88 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES males and females, as indicated by these figures : the right and the left-hand grips of the boy imbeciles surpass those of the girl imbeciles by 6.2 and 5.6 kg. ; for the adults the corresponding figures are 15.0 and 11.7. But the grips among the boy morons exceed those of the girl morons by 8.9 and 11.8 ; the corre- sponding figures for the adults are 16.3 and 17.3 kg. Analogous results, I believe, obtain among young and old normal children. In the imbecile group the difference is greater for the right than the left hand, while the reverse is the case among the morons. The sex-difference is thus less in the young and the intel- lectually inferior than in the more mature and intel- lectually superior. Another unmistakable fact is the superior strength in all the groups of the adults as compared with the children. In the general averages this superiority amounts to 7.5 and 7.7 kg. for the right and left hands, respectively; in the moron group the corre- sponding differences are 5.1 and 5.4, and in the imbe- cile group 7.9 and 8.4 kg. It thus appears that many epileptics grow physically stronger as they pass from childhood to adulthood. How long this increase continues among normal persons has not been deter- mined. Summarising, we may say : (1) The dynamometer test reveals significant sex and maturity differences, and differences between the right and left hands. The males are stronger than the females, whether epileptic or normal ; the adults and the morons among epileptics relatively more so than the children and the imbeciles. The absolute strength is greater for the adults and morons: the younger and inferior (lower grade) are relatively VARIATION OF MENTAL AND PHYSICAL TRAITS 89 nearer together than the older and intellectually superior. The left-hand grip is stronger with some epileptics, the right with others; but altogether the left hand is apparently slightly stronger, as seen most clearly in Graph VII. The index, 1.01%, is quite different from the normal index, from 91 to 96%. It is seen that the epileptics approximate a condition of ambidextrality, just as dull and feeble persons do. The latter have ''two left hands" (Binet and Yaschide). A positive correlation has also been asserted between dextrality and intel- lectual ability for normal children (Smedley). (2) There is an increase in the strength of grip with increasing B.-S. ages, but the increase from age to age is not entirely regular. Tliis indicates either that the B.-S. classification is not entirely correct, or that epileptics do not follow the rule obtaining among normal persons, who show an increase with each chronological age (Smedley), or that the irregu- larities may be due to the fewness of the patients in some ages. The gains with increasing age are greater for the boys than for the girls, for the chil- dren than for the adults, and for the right than for the left hand (although the difference between the low grade and the high grade subjects appears to be greater for the left hand). (3) Our results are in harmony with the finding for public school children, that the most intellectual are the strongest (Carman, Smedley, Schuyten). (4) Dynamometry is a valuable test for purposes of diagnosis ; its value for purposes of mental classi- fication and for testing the accuracy of intelligence scales needs to be determined with greater accuracy with normal persons. So important has the grip 90 EXPERIMENTAL STUDIES OB' MENTAL DEFECTIVES (hand grasp) of an individual been considered that Barr regards it as more indicative of capacity than the language test : hand grasp and mental grasp go hand in hand (Mental Defectives, 1910, p. 162).^ Ataxia graphic Sway. An examination of Table XXII and Graph VIII shows that there is diminution with increasing B.-S. age of the bodily sway, both lateral and antero-pos- terior, but that this diminution is by no means regu- lar from year to year. Nevertheless, if we examine the averages for the two groups, we find that the sway is less for the morons than for the imbeciles, with one exception (females, eyes shut). From Table XXIII, it appears that the group difference will vary from nearly zero to almost 20 mm., that the differ- ence in absolute terms is larger with the eyes shut than with the eyes open (one exception), that the difference for antero-posterior is larger than for the lateral sway, and that the sex-difference is only ap- parent with the eyes fixated. TABLE XXIII. Amounts by which the Ataxiagraphic Sway of Imbeciles exceeds that of Morons (based on Table XXII). r-Byea shut-^ . r-Eyes fixated— ^ A.-P. Lat. A.-P. Lat. Mm. Mm. Mm. Mm. Entire population 16.9 14.8 11.6 7.7 Males 13.5 12.0 9.1 21.9 Females 15.0 12.1 15.1 — .8 From Table XXIV, which gives the difference be- tween the absolute sway of the males and the fe- males (a positive number indicating that the sway *In a later contribution the dynamometry data will be tabulated ac- cording to chronological ages and compared with normal i>er£ormance8. CO CM rl "T H H W ftc tH 1-t CO T-H 00 00 CO CO rH T-l r-l C4 1-t k:^ to 00 LA lO lO CO CO ^ -M iH t- ^- rH rH CD C£> O U5 ^1 N fH ^ CC U5 CO CO iTi C^l U> CO m 05 1-1 (-1 M OJ CO CO 1-H -^ s —t M i-H O) 00 c- 00 o «o lO T-t ■< CO ^ c ■5 O <1 CO n ^; S o t- n 05 ■V <• . us u> I. S us t- US ffS TH ^ •♦ M 00 CO C-I CO 00 »« CO CO CO tH CO l-l in cvi f-< CO us CO tH 9> ec> '«' t> p w s > > > iS fc S >< X 5 2 92 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES is larger for the females ; a negative, larger for the males), it appears that, while the difference between the sexes is not very considerable in most cases, the sway is larger for the females than for the males. TABLE XXIV. Sex Differences in Ataxiagraphic Sway {based on Table XXII). ^Eyes shut-^ ^Eyes fixated-s A. -P. Lat. A.-P. Lat. Mm. Mm. Mm. Mm. Average of all ages — .5 4.9 2.2 2.6 Morons —1.1 3.4 —1.7 12.5 Explanations of signs above. The exceptions are the antero-posterior sway, eyes shut, both groups; and eyes fixated, morons. This seems to show that there is an ataxiagraphic sex- difference in epileptics associated with the lateral direction (larger relatively to the antero-posterior sway for females than for males). The exceptions can be gleaned from Table XXII. To summarize: (1) Apparently the ataxiagraphic sway varies with the sex, and with the eyes shut and open. It is slightly larger for female than male epileptics, par- ticularly in the lateral direction. The closing of the eyes accentuates the sway in both directions, more for the low than for the high grade cases. (2) The antero-posterior sway is usually consid- erably larger than the lateral, as is strikingly ap- parent in the Graph VIII, for eyes fixated. (3) The ataxiagrams are clearly smaller for the moron than for the imbecile group, so that there ap- pears to be a difference dependent upon the degree of intelligence. But the dependence is not very pre- cise, as the diminution of the sway from year to year is not very regular. (4) This is a valuable test for diagnosis in vari- VARIATION OF MENTAL AND PHYSICAL TRAITS 93 ous conditions, but before we are justified in using the test for purposes of intellectual classification or for checking intelligence scales it must first be dem- onstrated that the bodily sway among normal per- ,sons varies witli the degree of intelligence/ Analysis of the Mean Vaeiations. As a measure of the reliability of averages or the uniformity between the separate counts, whether averages or single determinations, which make up a series of determinations, we may use the mean vari- ation (M. v.). The reliability of the central tend- ency varies inversely with the size of the relative M. V. If the M. V. is large in relation to the size of the average {i. e., if the relative M. V., or coefficient of variability, is large), the average possesses little reliability. Likewise, if the measurements of a given trait from numbers of subjects grouped in the same mental age vary considerably, it follows that the test in question is not properly placed (not well attuned to its correct age) or that the variation is normally so large for the given trait that the test is worthless as a norm for a given age. It is clear that, in the very nature of the case, a norm is normative by reason of the fact that it indicates, within a certain range of variability, the expected performance for the age to which it is assigned. The larger the variability, the larger is the uncertainty of the norm; if the varia- bility is as large as the average, the latter, of course, fails utterly to represent any central tendency. At the same time, one must recognize that mental meas- ^In a later publication the ataxiagraphic data will be tabulated ac- cording to chronological age, and compared with the ataxiagrams for normal persons. 94 EXPERIMENTAL STUDIES OP MENTAL DEFECTIVES urement is concerned with the determination of tend- encies, not of absolute constants, and that a certain degree of variability of performance is thus per- fectly normal. The M. V. of a series of determinations is not very reliable unless the series contains a reasonable num- ber of observations. Because our number of sub- jects is small in some of the ages, it will be advisable to confine the analysis to the larger averages. The details may be left to the reader to glean from the tables. The coefficient of variability in the time required to name the four colors amounted to 28% for the en- tire population, 26% for the boys, 36% for the girls, 39% for the women, and 57% for the men. In the amount of irregularity the boys rank highest and the men lowest. The M. V. is larger for the imbeciles than for the morons. The variation is considerable in all the ages, and in Age VIII, in which the test is placed, it is even larger (37%) than for the general average. The coefficient of variability in the time needed to utter sixty ivords amounted to 25% for the entire population, 18% for the boys, 22% for the women, 28% for the girls, and 33% for the men. Here, again, the boys are the most regular and the men the least. The M. V. is considerable in most of the ages, and in Age XI, to which the test is assigned, it amounts to 24%, as compared with 13% for Age XII, the age to which it has been transferred in Binet's latest revision, and the age in which it is satisfactorily passed by our j3atients. The coefficient of variability in reading capacity, based on time, amounted to 28% for the entire popu- VARIATION OF MENTAL AND PHYSICAL TRAITS 95 lation and for the women, 26% for the girls, 29% for the boys, and 31% for the men. While the male sex appears to be the most variable, the differences are not very material. The amount of the variation dif- fers greatly in the different columns, but it is usually considerable in each age. In the ages in which the test is placed, VIII and IX, the coefficients are consid- erably larger than any of the above figures, namely, 41 and 50%, respectively. Even in Ages X, XI and XII, the coefficients amount to 60, 31 and 42%, so that while the average reading time diminishes with age, the variation among individuals continues large — a fact of common observation. That the amount of irregularity is partly due to the presence of a few very slow readings is indicated by the fact that the medians are uniformly and con- siderably smaller than the averages in Ages IX, X, XI and Xin. The coefficient of variability in memory capacity (memories from reading test) is 25% for the entire population, 15% for the boys, 24% for the girls, and 30% for the men and women. This points to a greater regularity, or homogeneity, of memory-capacity among the boys and a less regularity among the adults. In this respect the epileptic adults are obvi- ously much like the senile adults among normal per- sons. We have already seen that memory capacity among the children, as measured by the number of units retained, excels that of the adults. The size of the M. V.'s varies a great deal in the various ages and columns. The coefficient is 27% in Age VIII, in which the average number of memories is 3.7, instead of 2 as required by B.-S., and 27% in Age IX, in which there are 4.8 memories instead of six as re- 96 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES quired by the scale. This indicates (1) that the dif- ference in reproductive capacity between VIII and IX is not so large as indicated in the scale, at least for epileptics, and (2) that the variation for these ages is even larger than for the general average. On the other hand, the coefficients of variability in Ages XI and XII are only 17 and 19%, respectively. The variability in the test of motor performance — the form-hoard — amounted to 24% for the entire population, 25% for the women, 27% for the boys, 39% for the men, and 47% for the girls. Here the greater regularity in the boy and woman groups is manifest. The same truth appears if we confine the comparison to the morons : boys, 16% ; women, 16% ; men, 17%, and girls, 21%. The differences are slight, however. It is noticeable that the absolute M. V. is considerably smaller (at least in two cases) for the morons than for the imbeciles, and that the absolute M. V. decreases markedly with age. The exceptions are probably due to the fewness of the subjects. The variability is 14% for Age IX, 18% for Age X, 10% for Age XI, and 7% for Age XII. If we assume that the regularity of performance of a given activity increases with age, it appears that the patients are fairly well classified. The coefficients of variability for the right and left hand grips are shown in Table XXV. TABLE XXV. Coefficients of Variability in Hand Dynamometry. Right. Left. Entire population 20% 20% Boys 24 24 Girls 25 23 Men 13 15 Women 16 17 VARIATION OF MENTAL AND PHYSICAL TRAITS 97 It appears that the irregularity is about the same for the right and left hands, for the girls and boys, for the men and women, and considerably less for the adults than for the children. The coefficient is con- siderably less for the morons than for the imbeciles, the percentages for the right hand being, respec- tively, 16 and 25, and for the left hand 14 and 26. (Among normals the variability has been found greater in early adolescence than at any other time.) The relative M. V. 's for the body sway are shown in Table XXVI. TABLE XXVI. Coefficients of Variability in Ataxia^raphic Sway. r-Eyes shut-s r-Eyes fiiate(l-\ A.-P. L. A.-P. L. Entire population 24 29 24 34 Males 27 29 23 28 Females 21 30 33 34 It is seen that, without exception, there is among these patients less relative variation in the antero- posterior than in the lateral sway ; that the variation is not materially different, whether the eyes are closed or fixated, and that the differences between the males and the females do not seem to follow any rule. General Conclusions and Comparisons. 1. In most of the above tests there are maturity differences, that is, differences between the period of childhood and adulthood. The epileptic children excel in the color test (better average and lower M. v.), the 60-word test (better average and better gains), and the memory test (higher average and lower M. V.) ; and the adults excel in the reading test (better average), the form-board test (better 98 EXPERIMENTAL STUDIES OP MENTAL DEFECTIVES average), and the dynamometer test (better average and lower M. V.)- Accordingly, an adult who grades, say, X years mentally is not the same intellectually as a child with a X-year mind. The results, however, seem to show that the differences are not so large but that a common scale can be legitimately used for both children and adults — unless, indeed, the differ- ences are larger among normal than abnormal persons. Whether the adults attain a higher intellectual sta- tion cannot be determined reliably from the data in this chapter, but must be ascertained from the table of distribution in Chapter I. In tliis table it appears clearly that the adults reach a higher status : there are appreciably more child than adult idiots (3.7% more) and imbeciles (14.3%), but appreciably less child morons (11.8%) and normals, retardates or de- viates (5.3%). The inference from the mere curve of distribution, however, involves a certain source of error, because there are nine children chronologically less than thir- teen years of age who do not grade as feeble-minded by the B.-S. scale (that is, they are retarded less than three years). These children, therefore, belong to the group of normals or deviates, but could not be so classified because they failed to satisfy the thirteen- year standard. But it is probable that if due allow- ance is made for tliis error, the conclusion will still stand that the adults are intellectually superior to the children. It would thus appear that the average epileptic, in spite of the dementing tendency of the disease, makes intellectual progress from childhood to adulthood — ^unless, perchance, the onset of the at- tacks occurred late in life or unless the attacks were VARIATION OP MENTAL AND PHYSICAL TRAITS 99 more than ordinarily light in the case of this group of adults, both of which suppositions are improbable. From this fact we may conclude that institutions for epileptics should be made genuine training schools, not mere asylums. 2. There are likewise sex differences in most of these traits. The females excel in the color test (in average speed, although the M. V. is less for the boys), the reading test (average and M. V.), and the 60- word test (average, and gains for girls; but the men also have a high average), and the males excel in the form-board test (particularly in the average for the high-grade patients, and in the average and M. V. for the boy morons ; but the women also have a low M. v.), the memory test (average, and M. V. for boys), the dynamometer test (average, and also gains for boys), and the ataxiagraph test (average extent of sway). It is seen that the epileptic males excel in more tests than the epileptic women, and that the best single group is that of the moron boys. These results are confirmed by three other facts. First, by the table of distribution (Chapter I), which contains a larger percentage of high-grade males than females. Although there are slightly more male than female idiots (.7% more), the male imbe- ciles are fewer (by 7%), while the male morons (by 4.8%) and male retardates (1.4%) are more numer- ous. Likewise if we consider the separate columns, there are less men than women idiots (by 2.1%) and imbeciles (by 2.4%), but more men morons (by 3.1%) and retardates (by 1.3%) ; and there are more boy than girl idiots (by 5.7%), but decidedly less boy im- 100 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES beciles (by 20.8%) and decidedly more boy morons (11.5%) and retardates (3%). The obvious superi- ority of the boys is confirmed by the above figures, although a relatively large percentage of the idiots are boys. A second means of corroboration is to determine the relative station of the boys and girls by calculat- ing the number of years of retardation for each child, that is, the number of years between the child's chronological age and his B.-S. age. This could not very well be calculated for the adults, partly because the B.-S. scale does not extend beyond Age XIII (we have no other satisfactory tests for the higher ages that have yet been standardized), and partly because the attempt would involve a certain absurdity. Let us assume that two adults, respectively 30 and 40 years old, just grade XX years by a serial intelli- gence scale. We should not thereby be able to say that the 30-year-old is retarded 10 years intellec- tually in the same sense in which a 13-year-old grad- ing III years is retarded 10 years, since the increase in intellectual capacity from 20 to 30 would very probably not amount to 10 years. It would be less than the difference between 3 and 13. Nor, again, should we be able to say that our 40-year-old was retarded twice as much as our 30-year-old, for ordi- narily for the average person there is probably little increase in intellectual strength from 30 to 40 (there would, of course, be a great increase in erudition on the part of the studious). Hence, our 40-year-old would be retarded only slightly more than our 30- year-old. VARIATION OP MENTAL AND PHYSICAL TRAITS 101 Fnrthermoi'e, in restricting the following tabula- tion to the children, it is well to point out certain pos- sible errors. The amount of retardation is probably somewhat exaggerated in the case of those who pass the tests of the higher ages, as the norms are prob- ably too difficult. Nor, following the above reason- ing, does a unit of retardation mean exactly the same thing throughout the scale. The difference between, say, Ages II and III is larger than between Ages XII and XIII. A 20-year-old person who passes only Age XII, and a 12-year-old child who passes Age IV, would both be rated as eight years retarded, but the eight years from 4 to 12 almost certainly represent a larger difference than the eight years from 12 to 20. Moreover, the retardation is probably exagger- ated for the few who pass Age XIII, because these few might have gone higher, but there was no means of determining this by the scale. In Table XXVII the children are arranged in their chronological (actual) ages from 5 to 20, and the average amount of retardation is given in years for the patients of the same chronological ages. In Table XXVIII the amount of retardation is averaged for the children in each B.-S. age. The grouping in Table XXVII is based on chronological age, and in Table XXVIII on B.-S. ages. Here we see that the boys are retarded less than the girls: in Table XXVII by 0.8 of a year for the general average and by 0.9 for the average of ages 14 to 20 (with only two exceptions in the individual ages) ; and in Table XX VIII by one year for the gen- eral average (with three exceptions in the individual 102 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES O o «> I eo tH ** eo e<9 eo « <; t>oo it 00 w o '"' «D e >T) <1 a 13 t- oos iS "h "o •r< d W o oo" t)O0 - « a >> o o UJi J3 o o" M 55 *M T* ^^ rs I b^ ^ '"5 •^ US US 1-4 ^ 1— ( > s r >n M 4-» **k ^ IM 13 >> 0. E-H 006 • CQ 09 QQ l> rH a « g 4) li eS * as VARIATION OP MENTAL AND PHYSICAL TRAITS 103 ages). The superiority of the boys, whichever method of tabulation is considered, thus amounts, in round terms, to one year/ Finally, if we average the M. V.'s in all tests (ataxiagraphic excepted), we get the following re- sults : Boys. Girls. Children. Men. Women. Adults. Males. Females. 23.1 30.9 27.0 34.0 26.7 30.3 28.5 28.8% There is no difference between the males and fe- males, although the men are inferior and the boys distinctly superior. Altogether we may say, then, with confidence, that the males, and particularly the boys, in this group of epileptics, are superior to the females — a fact of common observation at the institution. Further study would determine whether this conclusion is valid for epileptics generally. The conclusion is ob- viously of interest, because, so far as concerns the children, one would expect the girls to surpass the ^Attention may be directed incidentally to a number of other facts which appear in the above tables. Thus the amount of retardation in general becomes greater as the child grows older, which is evident from the fact that the figures grow larger in the higher chronological ages in Table XXVII. This indicates a progressive loss of intelli- gence with increasing age, relative to the normal improvement, quite analogous to the progressive loss in manuometry exhibited by juvenile delinquents — unless, forsooth, this loss can be accounted for by the considerations advanced above. It seems natural that the normal child should outstrip the defective more and more, so that the gap between them becomes progressively larger. On the other hand, the loss with increasing B.-S. ages becomes less and less, the average for Ages I to VII (idiots and imbeciles) being 10.1 and for Ages VIII to XIII (morons and retardates), 5. The pro- gressive loss here is, no doubt, due to the fact that few children at the institution were under seven years of age. The feeble-minded status of the group is emphasized by the fact that the average amount of retardation is over 7 years. Yet there were 4 girls and 5 boys who were merely backward (retarded less than three years). 104 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES boys, since they mature earlier, and because it does not square with public-school statistics of retarda- tion, promotion, elimination and repetition. A study of fifteen city school systems showed that there were more retardates and repeaters among the boys than among the girls {137o more of each), and a larger percentage of elimination (17%) and non-promotion (Ayres). Either these findings do not apply to epi- leptic children, or otherwise our group has been spe- cially selected in some way. Apparently the disease (epilepsy) plays greater intellectual havoc with girls than with boys. On the other hand, our findings agree with the gen- eralization that there is a larger percentage of males at both extremes of the human scale — more very low- grade men (idiots) than low-grade women^ and more very high-grade men (geniuses) than high-grade women. The Germans, it is said, have a name for boy geniuses, but not for girl geniuses. Having these sex-differences in mind, a question of moment confronts us : Can we measure the intelli- gence of boys and girls accurately by the same scale? It would be unwise to attempt to answer this ques- tion on the basis of our results with epileptics. It is enough to say that our results emphasize the neces- sity of a thorough experimental study of the ques- tion. 3. The results indicate that the B.-S. scale is rather more satisfactory than appeared from the facts adduced in Chapter 11. "While it does not scale different grades of intelligence with the degree of accuracy which would be desired, it does enable us to grade and classify defective individuals far more rapidly and satisfactorily than would be possible VARIATION OP MENTAL AND PHYSICAL TRAITS 105 hy the ordinary methods of observation. The facts which perhaps tell more strongly against than for the accuracy of the scale are the averages in the color, dynamometry and ataxiagraphic tests, and the coefficients of variability. The latter range from 15% to 57%, and average almost 30%. Mental measurements, of course, are always variables, but the variations in our results seem to be rather too large. In many psychological measurements a varia- tion of from 10 to 15% is considered large. Perhaps we may fix 25% as the maximal permissible coefficient of variability. But we need to determine hy experi- mental means what should constitute a normal or maximal coefficient of variability for normal age- groups (normal persons). At the present time we are unable to assign any definite value to the age norm of variation (M. V. norm).^ No one, however, need be deterred from using the scale before this work has been done. Whatever its imperfections, it affords a practical, easily administered, objective, systematic method of grading defective children and adults, which, while not as yet maximally accurate, approximately locates the mental station of the indi- vidual. At the same time we must not blind our- ^Courtis tested 23 eighth-grade pupils, the survivals of eight years of "passing" on a 70% basis (and therefore presumably closely graded), and found the relative M. V.'s to amount to 18% in an addition speed test, "3% in age, 7% in handwriting, * * * \o% in formal Eng- lish grammar, 26% in height, 30% in memory of important dates and men, 60% in ability to reproduce after 24 hours the main points of a historical passage read once in class" (S. A. Courtis, Measurement of Groveth and Efficiency in Arithmetic, The Elementary School Teacher, II : 1911, p. 533). We shall be in no position to pass on the efficiency of the grading in graded schools until we have grade-norms of variabil- ity similar to the age-norms of which we have spoken above. In a serial investigation during the past year of various mental traits, I found surprisingly large M.V's, both in respect to the age and grade standards. 106 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES selves to the necessity of revising and improving the scale. 4. The data presented here will help us to a bet- ter understanding, not only of epileptic and other types of abnormal minds, but of normal persons as well, as soon as similar data are obtained from other surveys. Here we can take space for three brief comparisons only. According to B.-S., the time required to read the fifty-three words about a fire was as follows for the ages indicated : Age VIII, 45 sec. ; IX, 40 sec. ; X, 30 sec; and XI, 25 sec. Comparison with the corre- sponding ages in Table XI shows that the epileptics of VIII, IX and X required considerably more time (the times were 86.5, 61.9 and 44.6 sec, respectively), while the Xl-year-old epileptics required about the same amount of time — indeed, if we consider the medians, the time was actually shorter for the X- and Xl-year-old ejDileptics than for the normals. Here is probably a specific retardation effect incident to the disease (epilepsy). The epileptic children ac- quire the reading art more slowly than do normal children ; in fact, more slowly than the above figures indicate, because many of those who graded X and XI were chronologically from 12 or 13 to 18 or 20 years old. The following were the seconds consumed by the feeble-minded patients at Vineland in replacing the blocks in the form-hoard for each B.-S. age (accord- ing to figures supplied directly by Goddard) : II III IV V VI VII VIII IX X XI XII 150 90 70 40 35 29 20 19 19 17 15 It is seen from Table XIII that the feeble-minded reacted more rapidly than the epileptic in all ages up VARIATION OF MENTAL AND PHYSICAL TRAITS 107 to and including IX, in some ages very considerably faster, but that from Ages X to XII the epileptics slightly excelled. The higher-grade epileptics thus react pretty much like the higher-grade feeble- minded in the corresponding ages. It should be stated that the epileptics were tested when they were in their normal condition, and not in a state of stupor. In the latter condition the reactions would be markedly lengthened or would cease entirely. Finally, comparison for memory of digits may be made with Jacob's norms for London school children. He pronounced series of digits to about thirty pupils in each age from 8 to 19. The pupils wrote what they retained. The following averages are based on the highest number reproduced :^ Age. 8 9 10 11 12 13 14 15 16 17 18 19 Ave. 6.6 6.7 6.8 7.2 7.4 7.3 7.3 7.7 8.0 8.0 8.6 8.6 From these figures it appears that memory norms for digits cannot be established for each age. They seem to be approximately the same for several ages : 8 to 10 (6 digits), 11 to 15 (7 digits), and 16 to 19 (8 digits). It would not be profitable at this time to enter upon a discussion of age-norms for digits in the light of these and other available data. To es- tablish these norms upon a satisfactory basis far more children in each chronological age must be tested than has hitherto been done. 5. As may be seen by a cursory glance at the graphs, there is a continuity of variation in all the single traits tested, within the limits of the ages plotted, for epileptics. The variations cluster around *Cf. G. E. Johnson, "Contribution to the Psychology and Pedagogy of Feeble-Minded Children," Pedagogical Seminary, 3 : 1895, 268-273. 108 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES one type or mode (subject to certain irregularities in some of the traits), and it is entirely probable that this one-type pattern of distribution would not be altered were the records of normal subjects included. It is impossible to group these traits in epileptics into disparate classes, separated by intervening gaps : they vary continuously from lowest to high- est. Accordingly, the epileptic is not ' ' abnormal" in respect to these traits in the sense that he constitutes a distinct type, a species apart from ordinary indi- viduals. With all his exaggerated ''variation" or "individuality," he still shares with the rest of us a ' ' common humanity, ' ' the common humanity of the race. He is not essentially different in kind from the ordinary person. What appears to be a qualitative difference is probably an exaggerated quantitative difference, because the traits of the epileptic fall within the surfaces of distribution for the human kind. He occupies the lower extreme of the surface. (It should be said that I am leaving out of the ac- count the lowest-grade epileptics, who are not in- cluded in our "trait" curves, and that the curves may be considered only indirectly as distribution surfaces.) In conclusion : The great need of a simple, objective scale for in- tellectually grading defective school children and juvenile (also adult) delinquents, for measuring the extent of involution changes produced by dementing psychoses, and for classifying institutional cases of mental defectives of whatever kind, and the present conspicuous lack of a generally accepted or satisfac- tory method, is recognized on every hand. Neither the educator, the alienist, the criminologist nor the VARIATION OF MENTAL AND PHYSICAL TRAITS 109 judge have at their ready disposal a valid method for locating mental station and classifying individ- uals according to degree of mental defect. A graded diagnostic scale of intellectual development com- mends itself because of its great practical utility, even though less valuable for the purpose of exact scientific diagnosis. Because of the present consid- erable utility of the Binet- Simon scale, and its greater prospective utility, to all students who have to do with deviating individuals, no pains should be spared to ferret out its present imperfections, so that an improved scale may gradually be evolved. Until this has been done we may continue to use the present scale, confident that there is no other single instrument available which gives us a superior pre- liminary survey of a defective or mentally deviating individual — a conclusion already emphasized by the writer elsewhere.^ At the same time it is necessary to sound a note of warning: the function of the B.-S., or any other graded scale of intelligence, is to give us a prelimi- nary, and not a final survey or rating of the indi- vidual who may be tested. The B.-S. testing is not to be regarded as a finality, but merely as a point of departure for further diagnosis. It gives us the first, and not the final word about an individual's mental status. Once the individual's mental station has been preliminarily and roughly determined by a graded intelligence scale, there remains the more difficult task of making a detailed, expert diagnosis for each case and a diagnosis of each fundamental trait or capacity. *The New Clinical Psychology and the Psycho-Clinicist, Journal of Educational Psychology, 2 : 1911, 121 and 191. 110 EXPERIMENTAL STUDIES OF MENTAL DEFECTIVES It is folly to suppose that any ordinary grade teacher who has taken a summer course in clinical psychology, or that any professional expert whatso- ever who lacks technical training in psycho-clinical or psychiatric methodology, is prepared to mentally diagnose defective or deviating individuals, or to conduct psycho-clinical research, or to extend the boundaries of clinical psychology. It would be no more absurd to expect a teacher or preacher or law- yer or old-line experimental psychologist to diagnose human disease from reading a book on medical diag- nosis. The idea, unfortunately, seems rapidly to be gaining ground that anyone, be he grade teacher, introspective psychologist, practicing lawyer or gen- eral medical practitioner, is able to make psychologi- cal diagnoses by putting the child through a few stock psychological tests. Nothing is more preposterous. The ability to make psycho-clinical examinations and diagnoses requires as high an order of trained skill as the ability to make neurological or general physical examinations and diagnoses. All that can reason- ably be expected from grade teachers, or anyone who has not had considerable first-hand experience in the mental examination of various kinds of mental de- fectives, is that they acquire sufficient skill to give certain standard tests, like those of Binet and Simon, according to directions. It is visionary to suppose that a differential diagnosis of cases can be made by a few routine tests. The results of such testing do have one value, however: they contribute valuable preliminary data for a later individual differential diagnosis by an experienced mental examiner. Clin- ical psychology will more speedily become a genuine science of the behavior of individual deviating and VARIATION OP MENTAL AND PHYSICAL TRAITS 111 abnormal persons, and will develop a practically use- ful art of orthogenics, if we recognize at the outset that the problems of this science are extremely com- plex, complicated and baffling, and that they can be adequately handled only by trained experts. GRAPti n. B.-S.A^es 3 ^ S 6 7 S