^ HoUinger Corp. pH8.5 LB pi \ A. UNIVERSITY OF OREGON PUBLICATION VOL. I SEPTEMBER, 1920 NO. 7 A Study of the Mental, Pedagogical and Physical Development of the Pupils of the Junior Division of the University High School, Eugene, Oregon BY GILES MUEREL RUCH i^ «^^ C5 %12I0' Application made at the postoffice at Eugene, Oregon, for entrance as second-class matter UNIVERSITY OF OREGON PUBLICATION The University of Oregon Publication Series is oifered in exchange for the publications of learned societies and institutions, universities, and libraries. Address inquiries to MANAGER, UNIVERSITY PRESS, EUGENE, UNIVERSITY OF OREGON, EUGENE, OREGON. Vol. 1. No. 1. The Efficiency of Oregon Children in the Tool Sub- jects as Sho-wn by Standard Tests. By Chester Arthur Gregory, pp. 51. Tables. November, 1919 $1.00 No. 2. An Experimental Investigation of the Process of Choosing. By Rajnmond H. Wheeler, pp. 59. January, 1920. 1.00 No. 3. Earthquakes in Oregon. By Warren Dupre Smith. Reprint. Map. February, 1920 25 No. 4. The Foster-Child Fantasy. By Edmund S. Conklin. Reprint. March, 1920 25 No. 5. The Synaesthesia of a Blind Subject. By Raymond H. Wheeler, pp. 61. May, 1920 1.00 No. 6. Fossil Mollusks from the John Day Basin in Oregon. By G. Dallas Hanna 50 No. 7. A Study of the Mental, Pedagogical and Physical De- velopment of the Pupils of the Junior Division of the University High School, Eugene, Oregon. By Giles - Murrel Ruch 1.00 UNIVERSITY PRESS PUBLICATION COMMITTEE Eric W. Allen, Manager University Press, Chairman. F. G. Young W. F. G. Thacher E. L. Packard H. D. Sheldon A. R. Sweetser M. H. Douglass University PresIs 1920 .7Tf UNIVERSITY OF OREGON PUBLICATION Vol. I. September, 1920 No. 7 A STUDY OF THE MENTAL, PEDAGOGICAL AND PHYS- ICAL DEVELOPMENT OF THE PUPILS OF THE JUNIOR DIVISION OF THE UNIVERSITY HIGH SCHOOL, EUGENE, OREGON BY GILES MUREEL EUCH Principal of the University High School and Assistant Professor of Education University of Oregon OUTLINE Section I. Introductory. A. Statement of the Problem. B. Nature and Scope of the Investigation. Section II. Tests of General Intelligence. A. Individual. 1. Stanford Eevision of Binet-Simon Tests. B. Group Tests. 1. Army Group Examination Alpha. 2. Chicago Group Intelligence Test. C. Teachers ' Estimates. Section III. Acceleration, Eetardation, and School Progress. Section IV. Measures of School Attainment. A. School Marks. B. Standard Educational Tests. 1. Courtis Standard Research Tests in Arithmetic. 2. Kansas Silent Reading Tests. 3. Gregory Language Tests. 4. Ayres Spelling Lists. 5. Ayres Handwriting Scale. 6. Douglass Standard Diagnostic Algebra Tests. Section V. Anthropometric Measurements. A. Height. B. Weight. C. Lung Capacity. D. Strength of Grip. Section VI. Correlations, Discussion, and Conclusions. A. Reliability of Tests of Intelligence in Comparison with Teachers' Estimates. B. Use of Tests as a Basis for Special or Forced Promotions. C. Relation of Intelligence to Eetardation. D. Physical Factors in Relation to School Work. E. Correlations in School Abilities. F. Establishment of Just Standards for School Accomplishment. G. Conclusion. 1] study of Pupil Development SECTION I. INTRODUCTORY The Problem The purpose of this investigation is that of suggesting in several directions how the methods of the scientific study of educational problems through the medium of the various tests of general intelli- gence, pedagogical attainment, physical development, and the like, can be applied to the practical problems of school administration. The position is taken here that these types of objective measure- ments mentioned above, although admittedly as yet in the experi- mental stage, can furnish data of the greatest value in supplement- ing the traditional methods long in use as the basis for promotions, assignment of school marks, establishment of just standards of attainment, and many other similar problems of the administrator and teacher. It should be emphasized at the outset that it is not the purpose or spirit of this discussion to advocate the substitution of new and not thoroughly established practices for those methods which time has perfected and found valuable, but rather, to advo- cate the addition of these newer methods as supplementary sources of knowledge and thus make surer the validity of our school room practices. An imperfect tool is often better than no tool at all and a new tool of inferior design is often more serviceable than a once- excelleut old instrument. Wliat is needed is a judicious use of both the old and the new methods of educational practice. Nature and Scope of the Investigation The experimental work reported here is confined to the three grades of the junior divisioji of the University High School at Eugene, Oregon. These are the three years of the tj'pical junior high school under the 6-3-3 plan, i. e., extending from the seventh up to and including the ninth grades. The classifications dealing with age and grade and practically all of the tests were made dur- ing the second semester of the school year 1919-20. In a very few cases the data were obtained late in the year 1919. In the grades included there are about one hundred and twenty-five pupils rang- ing in age from less than twelve years to twenty years in one case. The pupils of the University' High School are practically an unselected group as far as entrance restrictions or preferences for I 2] University of Oregon High School admission are concerned. In accordance with an agreement entered into with the school authorities of the city schools of Eugene, the University High School draws from a certain section adjacent to the school in much the same way as do the other separate schools of the city. However, in spite of this arrangement, the students of the school represent a composite group composed of two distinct and very different elements. The first element comprises the pupils from the homes of people professionally connected with the State University. Since the school is an integral part of the School of Education of the University and is located on the campus in the building primarily used for the Department of Education, the University High School quite naturally draws heavily from faculty homes. It will be shown later that intellectually and culturally these pupils are above normal. The other element presents a sharp contrast in that this group of pupils represents homes that are socially and economically somewhat inferior, if anything, to those of certain other districts of the town. The parents of this second group are to a large extent day laborers and chiefly of the unskilled type. Although neither of these groups is numerically large, they do act as selective agencies which must later be taken into account in certain of the findings discussed in this paper. The tests given fall into three classes : 1. Tests of general intelligence. 2. Standard educational tests. 3. Physical or anthropometric measurements. In each case several different types of tests are given and these will be discussed under the appropriate headings in later sections of this study. The final section will deal with the attempt to cor- relate and classify these diverse measurements, to show their admin- istrative implications, to suggest problems which can be attacked by this method, and, finally, to draw certain conclusions apparent in the statistical treatments used here which are of interest to school administrators and teachers in the public schools, particularly in junior high schools. Specific acknowledgments for assistance in the work reported here are numerous since the detailed experimentation was done at all times in a co-operative way. Especially was the assistance of Professor H. R. Douglass of the Department of Education of the greatest value at all times, both in a material way and for helpful criticisms. Together with Mr. Peter L. Spencer, the former is chiefly responsible for the section on standard educational tests. [3] study of Pupil Development Miss Lexie Strachan, Mr, George E. Finnerty and the writer gave the individual Binet examinations. The first two mentioned worked under the direct supervision of Professor B. W. DeBusk of this department. Mr. Finnerty also assisted with the physical measure- ments for the boys and Professor Harriet W. Thomson of the Department of Physical Education for Women of the University kindly gave the services of her class in anthropometry for the physical measurements of the girls. Miss Strachan also checked some of the statistical treatments. The following members of the University High School staff made the estimates of intelligence quoted in Section III : Professor Douglass, Mrs. Geo. B. Goodall, Mrs. Edith B. Pattee, Mrs. Geo. S. Bendshadler, Mr. Victor P. Morris and Mr. Peter L. Spencer. SECTION II. TESTS OF GENERAL INTELLIGENCE For the purposes of this investigation it was finally decided that the Binet-Simon tests as revised and extended by Dr. L. M. Terman and his associates offered the highest degree of reliability of any single criterion of native ability or general intelligence. These tests have been carefully standardized and have the additional advantage of having been widely used upon California and eastern children. If marked sectional differences exist in the results of the use of intelligence tests, Oregon children would be likely to resemble California school pupils more closely than those of eastern states. The tests were in all cases given in a quiet room with only subject and examiner present. Care was taken to establish a good working "rapport" between the pupil and the examiner. The exam- iners were in all cases persons trained in psychological methods and the technique of intelligence testing. In cases of doubtful points in scoring all three examiners joined in arriving at an agreement. With few exceptions the responses of the pupils were recorded verbatim upon the regular test forms supplied by the author of the tests. The group tests used were two in number, the Army Group Examination Alpha and the Chicago Group Intelligence Test de- vised by Freeman and Rugg, In each case the exact directions fur- nished by the authors were followed without deviation. It was the original intention to use other group tests as supplementary and [41 University of Oregon High ScJiool corroborative data, but this was later given up because of the marked danger of introducing practice effects since there is a consid- erable degree of similarity between most of the group tests in use to- day. The sizes of the groups taking the tests varied from ten to sixty, but at no time was a larger number tested at any one time. A large well-lighted room with arm-rest chairs was used for the exam- inations. All of the group tests were given by the writer with one or more assistants present to supervise the work. It is interesting to record that not a single attempt at cheating was noted by the examiners in any of the tests. It is not thought that group tests do more than approximate the accuracy of the individual examinations but they do possess certain advantages from the administrative point of view which weigh heavily in actual school practice, viz., the great economy in time consumed in giving the tests. A group of one or even two hundred pupils may be tested in forty-five minutes with the Alpha test or in slightl}^ less than thirty minutes with the Freeman-Rugg scale. It would therefore be possible to test entire schools where the enroll- ment does not exceed the number stated above at one time. An hour 's time cannot well be held to be prohibitive, and there are few sources of information of vital significance which are more econom- ical in time or effort expended. More will be said later on upon the topic of the practical applications of intelligence ratings. Tables 1-10 and graph 1 summarize the various findings and are self-explanatory for the most part. Table 1 gives the classification of the degrees of intelligence as tentatively suggested by Professor L. M, Terman of Stanford University, TABLE 1 Terman 's Classification of the Intelligence Quotients* /. Q. Classification. Above 140 " Near ' ' genius or genius. 120-140 Very superior intelligence. 110-120 Superior intelligence. 90-110 Normal, or average intelligence. 80-90 Dullness, rarely classifiable as f eeble-mincledness. 70-80 Border-line deficiency, sometimes classifiable as dull- ness, often as feeble-mindedness. Below 70 Definite feeble-mindedness. In the tables of the distributions of the scores which follow, the scores fo? the Arm,y Alpha are in all cases the "raw" scores ; in the *Terman, L. M. : The Measurement of Intelligence, Houghton MifBin Co., Boston, 1916. p. 79. [5] Study of Pupil Development >. \u UAI __^ " — s w o (0 s ca 'A Mi !A !& '^ _ 1 L -_ 1 - - 1 . i - - - - 1 - - - - - 1 ~1 1 n n _! 1 1 ^ &-54-4 s_ t _ _ ^3_ 1 1 -t — 1 - •" tl^Mn i 1 :♦ - - - - ~ ^ ^1 o t 1 j4- oil s 1 r J" 1 !rt ri ~T 1 "~ t-' cr 3 ^ - - 1 1 e* frt^ 0. p 1 ~ 1 h- "^ ►J 3 3 " ~ ■~ la ^ > J 1 1 h 1 9 ~1 ■^ n _ _ _ _ s _ _ ^.i- 1-* . (fj -xi — ^ s Pf _ n r ""i f - 1 b J{ 1 1 . p 1 1 fe D _ 1 1 ^ 1 * — 6- 1 > — '«- j b ~ ! - — Boys 77.0 75.1 67-82 18 Girls 79.0 79.2 66-89 17 Both Sexes 78.0 77.2 66-89 35 IS I, II, and III Combined — Boys 80.0 79.9 67-93 59 Girls 83.0 81.8 65-95 61 Both sexes 82.0 80.9 65-95 120 20] University of Oregon High School TABLE 21 Showing the distribution of pupils failing or conditioned (i. e., an average semester grade of 74 or below) according to age-grade classification as acceler- ated, normal, and retarded. Accelerated Normal Betarded Total Number failed or conditioned 3 4 11 18 In percentages 16.7 22.2 61.1 100.0 Number in group 41 44 35 120 Per cent of group failed or conditioned 7.3 9.1 31.4 15.0 TABLE 22 Showing the distribution of pupils failing or conditioned (i. e., an average semester grade of 74 or below) according to the classification as "Superior" (L Q., Ill or above), "Average" (L Q., 90-110), and "Inferior" (L Q., 89 or below). ' ' Superior " " Average " "In ferior ' ' Total Number failed or conditioned 2 8 8 18 In percentages 11.2 44.4 44.4 100.0 Number in group 50 50 20 120 Per cent of group failed or conditioned 4.0 16.0 40.0 15.0 TABLE 23 Showing the median scores in the standard tests given in grade VII in comparison with the published norms. 1 Vumier a ttempted Numl >er correct Sc 'ore Tests Norm U. E. S. Norm V- H. S. Norm U. H. S. Courtis: Addition 10.5 7.0 6.5 4.0 Subtraction 11.5 8.0 8.5 7.0 Multiplication 10.0 6.0 6.5 3.5 Division 8.5 6.0 7.0 5.0 Kansas Silent Eeading Test 16.2 21.4 Gregory Language Test 23.S* 26.8 Ayres Spelling 73 75 Ayres Writing 57 48 ♦Gregory, C. A. : The Efficiency of Oj-egon School Children in the Tool Subjects, Uni- versity of Oregon Publication, 1, No. 1, 1919, p. 50. (Tentative norms based upon 1024 Oregon seventh grade pupils representing 50 schools. ) 21] study of Pupil Development TABLE 24 Showing the median scores in standard tests given in grade VIII in com- parison with the published norms. Number attempted Number correct Score Norm U-H.S. Norm U.H.S. Norm U.H.S. Courtis: Addition 12.0 9.0 8.0 6.0 Subtraction 12.5 9.0 10.0 7.0 Multiplication 11.5 8.0 8.0 6.0 Division 10.5 7.0 9.0 6.0 Kansas Silent Reading Test 19.2 25.2 Gregory Language Test 30.7* 36.3 Ayres Spelling 84 93 Ayres Writing 54 ♦Gregory, loc. cit., p. 50 (Tentative norms based on tests of 1029 eighth grade Oregon school children from 50 schools ) . TABLE 25 Showing the median scores in standard tests for the ninth grade in com- parison with the norms. Test Norm Test Ayres Spelling .... 100 Ayres Writing .... 64 Douglass Algebra 29.6* 30 *DouGLASS, H. R. : A Series of Standardized Diagnostic Tests for the Fundamentals of First Year Algebra. An unpublished thesis submitted to the Graduate School of the Uni- versity of Missouri, 1920. TABLE 26 Showing the medians and ranges of scores made in the Kansas Silent Read- ing Tests for groups of accelerated, at age, and retarded pupils. Grade VII Grade VUI Median Bange Median Eange Accelerated 20.5 17.8-27.7 29.4 15.4-40.4 At age 21.1 10.5-38.9 32.0 14.9-54.8 Retarded 20.3 17.0-32.9 19.1 5.6-39.8 TABLE 27 Showing the m.edians and ranges of scores in the Kansas Silent Reading Tests for groups of Superior, Average, and Inferior pupils. Grade VII Grade VIII Median Eange Median Eange Superior 32.0 18.0-39.0 35.0 17.0-45.0 Normal 20.0 11.0-33.0 22.0 11.0-40.0 Inferior 17.0 13.0-25.0 24.0 10.0-29.0 [22] University of Oregon High School TABLE 28 Showing the medians and ranges of scores (number worked correctly) for the Courtis Research Tests in Arithmetic by groups of accelerated, at age, and retarded pupils. The scores are the sums of the separate scores in the four fundamental operations : Accelerated At age Eetarded Grade VII Median Bange 17.0 4.0-31.0 21.0 10.0-34.0 12.0 5.0-23.0 Grade VIII Median Bange 25.5 17.0-46.0 29.0 11.0-57.0 20.5 11.0-50.0 TABLE 29 Showing the medians and the ranges of scores (number problems worked correctly) for the Courtis Research Tests in Arithmetic by groups of Superior, Normal, and Inferior pupils. The scores are the sums of the separate scores in the four fundamental operations. Grade VII Median Bange Superior 20.0 4.0-34.0 Normal 20.0 12.0-27.0 Inferior 12.0 5.0-26.0 Grade VIII Median Bange 29.5 11.0-46.0 23.0 10.0-57.0 20.0 11.0-22.0 TABLE 30 Showing the medians and ranges of scores made in the Gregory Language Tests by groups of pupils classified as accelerated, at age, and retarded. Accelerated At age Retarded Grade VII Median Bange 34.3 23.2-52.6 20.0 5.1-61.8 22.1 7.2-46.7 Grade VIII Median Bange 44.2 18.8-61.4 32.0 14.9-54.9 28.2 10.3-46.0 TABLE 31 Showing the medians and ranges of scores made in the Gregory Language Tests by groups of pupils classified as Superior, Normal, and Inferior. Grade VII Grade VIII Median Bange Median Bange Superior 34.3 19.4-61.8 43.4 14.9-61.4 Normal 32.8 5.1-46.7 31.9 10.3-54.8 Inferior 23.8 12.7-28.8 18.5 7.2-23.6 TABLE 32 Showing the results in the Stone Reasoning Tests in grades seven and eight. Total Score score per Number Number for WO Median problems problems Per cent Grade No. class pupils score attempted right accuracy VII-A 36 300.9 836 9.1 364 212 58.2 VIII-B 21 177.1 843 9.2 194 128 66.0 VIILA 17 183.4 1,079 10.8 169 126 74.5 [23] study of Pupil Development SECTION V. ANTHROPOMETRIC MEASUREMENTS Justification for the inclusion of the results of physical measure- ments may seem necessary in view of the fact that such data has not generally proved to be of the highest value. However, the close relationship between the physical and mental development of the human body cannot well be doubted regardless of whether such parallel maturation is correlated by direct causal factors or whether it represents merely concomitant relationship. For this reason, it can be shown that knowledge of certain physical factors does contribute data which will prove to be of considerable value indirectly in the understanding of individual cases among school children. Phj^sical immaturity, for example, is a frequent factor in low scholarship as can be shown from the experience of any observ- ing teacher. The measurements selected for use in this study include : 1. Height standing. 2. Weight. 3. Lung Capacity ("Vital Capacity- "). 4. Strength of Grip. The weights were all taken without clothing and were recorded in pounds and fractions. Height was likewise read to the nearest tenth of an inch. Both standing and sitting height were obtained although only the results in the first are used here. Lung capacity was measured by means of the wet spirometer and the strength of grip by means of the Smedley combined dynamometer and djTiamo- graph. In using both of the last two instruments three, or in a few cases more, trials were allowed and the one best record preserved. Likewise in the lung and strength measurements the work was done with small groups in mild competition in order to insure a high amount of interest and effort. The dynamometer was held in all of the tests at the side of the leg of the subject without touching the subject. Only the reading for the right hand will be considered here. The matter of the choice of the norms to be used as a basis for the comparisons presented considerable difficulty due to the fact that the exact conditions under which the various investigators have worked have not been standardized. In the matter of weight, for example, most of the studies have been made with clothing and [24] University of Oregon High School corrections for this factor are at best but approximate. It was finally decided here that for weight and height the results of Baldwin* were best suited to our purposes, first because of the fact that his weight norms are taken without clothing, and secondly, because his study was primarily made in university high schools or other demonstrational schools which are more likely to resemble the University of Oregon High School than are the results from studies of city school children generally. For vital capacity and strength of grip the norms of Smedleyt will be used. Tables 33 to 34 show the heights and weights of the University High School children in comparison with the Baldwin data. In general it will be seen that the Oregon pupils resemble those of the University of Chicago schools and the Francis W. Parker School very closely. Where our cases are sufficiently numerous to permit of comparisons the deviations in height or weight for either boys or girls are not significant in amount or in a constant direction. The western boys will average somewhat shorter than the eastern pupils and the Oregon girls slightly taller. For weight the Oregon boys show small but definite superiority, but in the case of the girls, this is not true to any extent. It should be pointed out that Baldwin's pupils are physically a superior lot in comparison with school chil- dren in general. He states : " the children have been under school medical inspection, directed play, and physical education. That these factors are important educational agencies is shown by the fact that, on the average, these children from the Horace Mann School, the University of Chicago Elementary and High School, and the Francis W. Parker School are taller and heavier than any other group of children so far recorded among over a million studied. ' '$ The differences between the results obtained for vital capacity of the Oregon pupils and the norms of Smedley afford the most striking single fact in the data on physical measurements. Both boys and girls (Table 37) show markedly greater development of lung capacity at all ages. If the vital index as often computed by the formula lung capacity V. I.= weight is obtained for each pupil, it will be found that the medians for ♦Baldwin, B. T. : Physical Growth and School Progress. Bulletin of the U. S. Bureau of Education, 1914, No. 10, 1-215. tSMEDLEY, F. : Report of Department of Child Study and Pedagogic Investigations, 46th Annual Report of the Board of Education, Chicago, 1899-1900. Jloc. cit., p. 9. 25] Study of Pupil Development such indices at all ages are higher than norms obtained by dividing the birthday norms of weight and lung capacity furnished by Smedley (see Table 38). It should be remembered that Smedley's data for weight includes clothing and hence these computed indices will run slightly too low. This, of course, will far from explain the marked differences evidenced by the Oregon children. In treating data for such small numbers of cases as are con- cerned in this study, it is very difficult to make reliable comparisons with the published norms because of the influence of extreme vari- ates upon the measures of central tendency. For example, if our 125 pupils are classified according to ages, there will be in no case more than 17 pupils at any age and but one at certain ages. Moreover in certain cases, e. g., for the vital index, which is a fluc- tuating value for the different ages because of the imperfect paral- lelism in the increases of lung capacity and weight with age, it is difficult to compare such indices with any of the measures of intelli- gence like the mental age and the intelligence quotients. Mental age is a score in contrast with the I. Q. which is an index, but on the other hand, the I. Q. is believed to approximate a constant value for an individual year after j-ear. To correlate a somewhat constant index like the I. Q. with a fluctuating index like the vital index is unreliable if the pupils are of different ages. In order to make possible these comparisons of groups of pupils of differing ages, it is suggested that the vital index be compared with a theoretical vital index to be obtained by dividing the lung capacity norm by the weight norm at each age. The result will be a value ranging above or below 1.00 in much the same way as does the I. Q. Such an index might be termed an ''index of normality" and could be ap- plied in all cases where values which are statistically indices, in con- trast with scores, are concerned. Likewise the index obtained for strength of grip by a similar process : Grip (actual) ' ' Index of normality ' ' f oi' strength of grip = Grip (norm.) Tables 37 to 38 show the results of such correlation of these indices with the intelligence quotients. It is to be noted that the relation- ship between the vital index and intelligence which Smedley, De- Busk*, and others have postulated finds some confirmation in the correlations stated in Table 41. The chief limitations apparent in *DeBusk, B. W. : The Vital Index in Development, Ped. Sem., 24, 1917, pp. 1-18. [26 1 University of Oregon High ScJiool the use of this method would seem to center about the fact that there must be available reliable norms for any trait under consid- eration. In such measurements as physical development, there ap- pear to be regional differences which act as disturbing factors in the computations. However, the advantage of increased numbers with a corresponding increase in reliability of the statistical measures found should be considered as an argument in favor of this manner of treatment. At any rate it might prove valuable to try out this plan of correlation with other and larger groups of pupils and for other physical and mental traits where norms are to be had. The sexes need not be separated. Table 40 shows that strength of grip for the right hand shows a correlation of 0.35 for boys and —.08 for girls by the Pearson for- mula. In the same way the coefficients for the correlation of the 1. Q. and the "index of normality" for the vital index are 0.244 for girls and 0.240 for boys. The coefficients support the conten- tion of DeBusk and others that there exists a positive relationship between the vital index and mental ability. This relationship is not very perfect and is very probably of indirect nature. It does suggest the interesting possibility of physiological effects upon the mental processes by the efficiency of oxygenation through the medium of the blood stream. The ratio of oxygen intake to the weight of the tissues of the body could conceivably be a factor of influence upon the metabolism of the organism. The most important difficulty in establishing such a relation seems to lie in the fact that the spirome- ter readings are not exact measures of lung efficiency since different types of breathers and different amounts of effort in using the instrument cause different relations between the tidal, comple- mental, and supplemental air volumes to arise. There seems to be no method of regulating and standardizing these relations. In the- ory, these three volumes are summated to form the vital capacity. Tables 35, 36, and 39 show the comparative development of the pupils of the University High School and about 850-900 pupils of the city schools of Eugene in height, weight, and vital indices. The figures for the Eugene schools were kindly furnished by Superin- tendent W. R. Rutherford and are published with his permission. The advantages, in the main, are to be credited to the University school, but it should be remembered that the data for the city schools does not extend above the eighth grade. This would tend to cause the comparisons for the upper years to be made between pupils normal to the high school grades in the one case and in the [27] Study of Pupil Development other for over-age pupils of high school age who are still in the elementary schools. These comparisons are interesting again in that the larger num- bers of the city schools dispose of the possibility that the better physical development of the University High School pupils in com- parison with Baldwin's and Smedley's norms in a chance effect of the small number of cases. One possibilitj^ which should be called to attention in correla- tions between mental and physical traits is that such might be due merely to the varying degrees of effort expended by different types of individuals ; the more intelligent, perhaps, making greater efforts in the course of taking such measurements as grip and lung capa- city and hence introducing a type of spurious correlation. TABLE 33 Comparing the median heights of the pupils of the University High School with the medians for the pupils of the University of Chicago High School and the Francis W. Parker School (Baldwin). Heiglit in inches (medians) Boys — Age Oregon Chicago Diference N iiy. 52.7 55.9 —3.2 2 12y2 57.9 57.9 0.0 14 13 Vo 59.0 59.8 —0.8 17 14Vo 61.6 62.6 —1.0 12 151/. 63.3 64.7 —1.4 6 161/2 69.5 65.5 -f4.0 1 171.^. 64.5 66.5 —2.0 3 Girl's— iiy. 58.5 56.6 -fl.9 3 121/, 60.0 58.1 + 1.9 14 131/, 61.1 60.9 +0.2 17 141/, 62.2 62.0 +0.2 12 15 yo 61.9 62.6 —0.7 8 161/. 61.8 63.0 — 1.2 4 17V2 62.3 63.4 TABLE 34 — 1.1 1 Comparing the median weights of the pupils of the University of Oregon High School with the medians for the University of Chicago High School and the Francis W. Parker School (Baldwin). Weight in pounds (medians) Age Oregon Chicago Diference N Boys — 111/, 68.5 72.7 —4.2 2 121/' 87.5 77.8 +9.7 14 131/, 89.0 87.0 +2.0 17 141/, 100.0 102.5 —2.5 12 i5y; 109.5 108.5 + 1.0 6 16V, 144.0 114.5 +29.5 17y2 129.5 128.5 +1.0 [28] University of Oregon High School Girls— 111/2 95.7 72.5 + 28.2 3 121/, 80.5 83.0 —2.5 9 131/2 99.5 95.0 +4.5 17 141/2 103.0 103.5 —0.5 12 15 Va 109.4 111.2 —1.8 8 leVa 123.0 112.5 + 10.5 5 171/2 120.0 116.5 +3.5 1 TABLE 35 Comparing the heights of the pupils of the University High School with those of the public schools of Eugene according to the measurements furnished by Superintendent W. E. Eutherford of the city schools. Measurements for the city schools are averages, for U. H. S., medians. Boys Girls Age U. H. S. City Schools U. H. S. City Schools 11% 52.7 54.44 58.5 56.07 121/2 57.9 56.16 60.0 56.99 131/2 59.0 59.31 61.1 60.35 141/0 61.6 61.13 62.2 62.23 151/2 63.3 63.73 61.9 62.84 16 V> 69.5 67.40 61.8 62.63 i7y2 64.5 62.3 TABLE 36 Comparing the weights of the two groups of pupils in the same way as in the preceding table of heights. Boys Girls Age U. E. S. City Schools U. H. S. City Schools 111/0 68.5 73.98 95.7 74.31 121/, 87.5 79.94 80.5 81.32 131/, 89.0 87.16 99.5 97.43 141/, 100.0 97.29 103.0 104.00 151/0 109.5 114.08 109.4 109.77 I6I/2 144.0 128.07 123.0 115.61 i7y2 129.5 120.0 TABLE 37 Comparing the medians for Adtal capacity for the University of Oregon High School with those of Smedley. Vital Capacity Smedley 's Age in ( juMc inche 's Data Difference ^ Boys— iiyo 120.5 110.2 + 10.3 2 12^ 151.0 119.6 +31.4 13 131/0 164.0 137.7 +26.3 16 i4y. 170.5 154.5 + 16.0 12 15% 241.0 174.5 + 76.5 5 161A 237.0 206.2 +30.8 2 1714 209.0 218.8 — 9.8 1 Girl's— 11 y2 138.0 97.1 +40.9 3 12y2 150.0 105.9 +44.1 9 i3y2 164.0 116.9 +47.1 17 14% 152.0 128.8 + 23.2 12 15% 160.0 13.5.7 + 24.3 8 16% 197.0 140.6 +56.4 5 17% 196.0 142.9 +53.1 1 [29 study of Pupil Development TABLE 38 Comparing the vital indices of the University High School pupils with those computed from the birthday norms of Smedley. Parentheses show numbers measured. Boys Girls Age Oregon Smedley Oregon Smedley 111/2 1.75 ( 2) 1.53 1.44 ( 3) 1.36 121/2 1.90 (13) 1.52 1.87 ( 9) 1.32 131/2 1.90 (16) 1.54 1.65 (17) 1.28 141/2 1.73 (12) 1.56 1.48 (12) 1.24 151/0 2.01 ( 5) 1.55 1.46 ( 8) 1.25 161/2 1.76 ( 2) 1.69 1.60 ( 4) 1.23 171/2 1.54 ( 1) 1.67 1.63 ( 1) 1.22 TABLE 39 Comparing the vital indices for the pupils of the University High School with data obtained for the city schools of Eugene by Superintendent W. R. Rutherford. Data for Eugene schools computed as averages, for the U. H. S., medians. For numbers of cases see preceding table. Numbers for city schools about 800. Age V. H. S. City Schools U. H. S. City Schools 111/2 1.75 1.554 1.44 1.386 121/, 1.90 1.553 1.87 1.382 13% 1.90 1.542 1.65 1.339 141/2 1.73 1.590 1.48 1.377 ]5i/o 2.01 1.517 1.46 1.366 161/2 1.76 1.676 1.60 1.367 TABLE 40 Showing the Pearson coefficient of correlation between the I. Q. 's and the "indices of normality" for strength of grip in the right hand (upon the basis of Smedley 's data). r P. E. N Boys +0.350 .078 57 Girls —0.080 .087 60 TABLE 41 Showing the Pearson coefficient of correlation between the I. Q. 's and the indices of "normality" for the vital indices computed from the data of Smedley. r P.E. N Boys +0.240 .089 51 Girls +0.244 .089 51 TABLE 42 Showing the medians for strength of grip in the right hand for boys and girls compared with the norms of Smedley. Measurements in kilograms. Boys Girls Age Oregon Smedley Oregon Smedley 111/2 24.25 20.03 19.00 17.65 121/2 27.50 22.45 23.00 20.19 131/2 30.00 26.43 25.00 23.49 141/0 33.00 30.40 25.50 26.10 151/2 38.00 36.38 26.00 27.91 I61/2 44.50 42.35 28.25 29.50 171/2 47.50 47.14 36.00 29.63 [30] University of Oregon High School SECTION VI. CORRELATIONS, DISCUSSION AND CONCLUSIONS In this concluding section the effort will be made to correlate the data presented before and make certain applications of our findings to school problems and at the same time suggest other problems which may be attacked by these methods. It is not be- lieved that the problems to be discussed here can be answered in any degree of finality, but it is thought that solutions may be suggested. At the outset it may be of value to raise the entire question of intelligence testing and other forms of educational measurements as well. In spite of the wide-spread use of such objective methods as those of the tests of native ability, it may fairly be said that among the rank and file of public school men and women there are still those who raise certain questions. They often hold to the view that careful and experienced teachers are just as accurate in their estimates of intelligence as tests can be made to be, or again, they ask: Even after one has found the mental age, the I. Q., or the score in Army Alpha, of what value is this knowledge? Are not such measures prohibitive in their demands upon time and energy? The last will be taken up first since it can be answered quite briefly. To take a concrete illustration, a pupil spends in a given grade in the course of a full year 's work of nine months something more than a thousand hours under the direct observation of the teacher, granting a six-hour day. It requires approximately one hour of these one thousand to give an individual Binet examination — or one-tenth of one per cent of the school year. If there are thirty pupils in a room and a group scale requiring an hour's time to give is used, the time expenditure is about two minutes per pupil plus about five minutes per pupil to score the results. If the results of such tests of intelligence have even the smallest practical value, the time expenditure is almost negligible. Every year we devote dozens of times this amount of time to one or another worthy cause which may be very much less closely related to school affairs. Pass- ing from this question to that of the possibility of intelligence tests being entirely superfluous since capable teachers of long experience can estimate ability with a reliability which approaches that of the tests, it should be recalled that it is true that the teachers of the University High School were able to estimate intelligence with a reliability of from .60 to .70 or more as shown by the coefficients of [31] iitudy of Pupil Development correlation. In view of the extremely favorable circumstances under which these ratings were made, it is doubtful whether much greater accuracy can be ordinarily obtained. If, then, we examine a few of the cases of displacement we will find ourselves forced to the conclusion that, although the reliability of the method of esti- mation is high in the mass, in individual cases it will lead to serious error. Stated in other words the method breaks down in just those individual cases where it could prove of the greatest value. To cite some illustrative cases, the following will serve the purpose : Case 1.— V. L., girl. C. A., 11-8. M. A., 15-5. I. Q., 132. Army Alpha, 83. Chicago, 49%. Average grade, 79. Teachers' estimate, 4.25 (pooled estimates of four teachers; range of estimates 2.0-6.0. See preceding section for basis of assigning ranks). This is a case of a quiet, retiring pupil, physically immature, somewhat irregular in attendance, and handicapped by ill health to some extent. Previous school advantages probably not of the best. The teachers have evidently judged her ability by her grades overlooking the above factors and the additional one that she is almost two years ahead of her grade and is thus subjected to ex- treme competition. Her real ability demands that she be rated 1.0 or a fraction more in the scale used for estimation. Case 2.— E. F., boy. C. A., 12-7. M. A., 16-7. I. Q., 132. Alpha, 133. Chicago, 47%. Average grade 73 (failure). Teachers' estimate 3.0 (pooled results for four teachers; range 1.0 to 6.0). Pupil vivacious, troublesome in minor ways, noisy, lazy, inattentive, self- satisfied, and more or less indifferent to praise or criticism. Could do the high- est type of school work if he so desired. Father an able professional man. Another case of teachers basing judgments upon school work. Pupil acceler- ated two years but bids fair to lose this position unless some method of inter- esting him in school work is found. Should be rated 1.0. Case 3.— C. W., girl. C. A., 13-10. M. A., 12-9. I. Q., 92. Alpha, 69. Chicago, 27%. Teachers' estimate 3.25 (pooled results for four teachers; range 2.0-6.0). Average grade 87. Pupil active, vivacious, energetic, works carefully and consistently, takes prominent part in school activities, a natural leader among her class-mates, and would be popularly described as ' ' looking intelligent. ' ' Her pleasant manners and excellent work habits cause her to be rated 3.25 when her real ability is 5.0 or 6.0 on our scale. She is just at grade for her age. Case 4.— M. G., boy. C. A., 15-9. M. A., 18-1. I. Q., 115. Alpha, 130. Chicago, 64. Average grade 73 (failure). Teachers' estimate 6.0 (pooled results of five teachers; range 4-9). One of the notoriously lazy boys of the school. Constant minor offender against school rules. Not responsive to social j^ressure. Prefers to read wild west stories to school books. Could easily do 90 per cent school work. Eated too low because of work habits. Should be rated 2.0-3.0. Is retarded about one yigar. Has failed before. If space permitted many other cases could be cited, e. g., the girl standing fourth in the entire school upon the basis of the I. Q. and beyond doubt to be rated at 1.0 on our scale was assigned the rank 6.0 by her teachers, the estimates ranging from 4.0-8.0. It will [32] University of Oregon High School be argued that these are the exceptions and not the rule. This is just the point in the position taken here. It is in the exceptional case that the intelligence examination proves its worth. We have shown by table 10 of Section II that 18.9 per cent of the pupils are displaced more than two groups — an amount which as was shown is a serious displacement — and that but one pupil in six can be located in the correct group as determined by the tests. Sixty-one and one- tenth per cent are located with a displacement of no more than one group — an amount which is not serious. This leaves one pupil in five who cannot be rated by the method of estimation because of complicating factors and it is just this pupil that most needs to be given an evaluation in terms of native ability. For these reasons it does not seem likely that the method of estimates promises any serious rivalry to the method of tests. Two other points suggest themselves as being worthy of brief mention. The first is that the method of estimates is almost as expensive in time and efforts as are individual tests and much more so than group examinations, as was shown by the fact that in order to obtain the high reliability of our estimates, some teachers devoted as much as five hours to the task. The second point is that both methods are of practical as well as instructive value since they are mutually corroborative. The fact that comparisons of the ratings obtained by the two methods show discrepancies leads to deeper analysis of pupil's abilities by teachers and a truer understanding of the complex relationships between the many factors combining to determine school attainment. The remaining point as to what are the schoolroom applications of the mental age, I. Q., or other intelligence ratings will not be further discussed here but taken up incidentally along with the other problems in the pages to follow. The next application of the method of objective measurements in school practice is that of the possibility of using standard educa- tional tests and intelligence scores as a basis for promotions, grad- ing, and school advancement. This is a new field where generaliza- tions are dangerous but offering such possibilities that a few ideas merit some consideration. In Section IV it was shown that the pupil who is considered as accelerated by the usual standards would in most cases be classified as retarded in an age-grade table based upon the use of mental ages. This fact has been repeatedly empha- sized by Terman and others and apparently passes unchallenged. Could, then, a pupil do the work in the grade which is normal for his mental age in a satisfactory manner and prove able to stand the [33 1 C)tudy of Pupil Development competition of pupils actuallj^ one or more years older? Obviously to transplant the twelve-year-old boy or girl testing at a mental age of fifteen (I. Q. approximately 125) into the first year of the high school without completing the work of grades seven and eight would be an entirely unwarranted educational policy which might entail no end of objectionable consequences. This hiatus in the fundamentals of subject matter would persist throughout the school and real life of the pupil and worse still the foundations for future building would be undermined. Such practices are not to be seri- ously considered. But, on the other hand, there may be an alterna- tive. As every teacher knows we have with us in the classroom a considerable group of pupils of superior ability who appear to be bored with the work of the school and grade and who in the end fall either into attitudes of apathy or petty mischief. It has been a rather frequent criticism of the schools of democracies that they look first to the average or sub-average child often to the neglect of real ability. This may explain in large part the fact that so many men of genius have been overlooked by their teachers only to burst forth like meteors in their later lives. There is little doubt that the lock-step method of school progress has tended to divert superior ability and energy, particularly in boys, into wasteful and negative channels. These criticisms are not intended to serve as arguments for falling into the errors of some of the European school systems of focusing the entire effort and attention of the schools upon the few of superior endowments at the expense of the many. One extreme is as bad as the other, except that the latter does seem to produce a certain body of social, political, and scientific leaders. The real issue is that any system which discriminates between the various levels of ability is wasteful and unfair. Just what the mechanics of a system permitting pupils to progress at rates deter- mined hy their individual abilities will prove to be is not within the scope of this discussion to suggest. Several efforts have been made in the past in the form of the Pueblo plan, modified monitorial sys- tems, the abolition of class instruction with a return to individual teaching, the Batavia scheme, and many others. Each of these has advantages and none are free from objections. Divergent as they are, the.y are in agreement on the one point that the present lock- step system is undesirable in view of the new knowledge of the magnitude of individual differences. Under present conditions we find in extreme cases where the pupils are too mature to derive full value from a grade that we are [34] University of Oregon High School forced to promote these ahead a half grade or a grade without the intervening work. That such promotions are make-shifts would not be denied. However, if under the present conditions forced pro- motions are necessary at times in individual cases or for administra- tive reasons, we should supplement our judgments in these promo- tions with whatever other types of data are at hand. That the results of intelligence and standard educational tests can be made to serve this purpose is well illustrated by the following case reported by the writer in Educational Administration and Supervision for February, 1920.* Here because of crowded conditions and unequal distributions of pupils it was found necessary to promote about ten pupils from the seventh B to seventh A section. It was the plan to select about that number from the lower grade upon the basis of ability, choosing the highest in the class. Teacher's estimates and judgments were unavailable since the pupils were all new to the school and had come from at least a half-dozen different schools, in some cases outside the state. It was finally decided to base the promotions upon the showing made in a "battery" of educational tests consisting of the Courtis, Kelly and Gregory tests. If time had permitted the Binet tests would have been given. Unfortu- nately these had to be deferred until later ; the results are, however, included in the table below, which is reproduced from the original paper. The pupils standing highest in the three tests were ad- vanced into the next section except for two rejections, one for physical immaturity and the other on account of parental objection. TABLE 43 Records of seven pupils who were advanced from 7B to 7A. Individual Tests Average Group Tests Pupil C. A. M. A. I. Q. Grades Chicago Alpha "A. H. 13-3 15-9 119 85 54 98 J. K. 12-9 14-9 114 80 36 114 H. C. 14-2 13-0 93 76 15 65 J. B. 12-4 14-0 114 82 451/.'. io9 K. E. 12-9 14-11 117 82 34 102 B. B. 11-8 18-3 113 85 38 97 E. T. 12-2 15-9 130 83 45 94 Averages 12-8.7 15-5.6 114.3 81.9 38.2 97.0 Medians 12-9.0 14-7.0 114.0 82.0 38.0 98.0 Records of pupils regularly belonging in grade 7A — C. A. M. A. I. Q. Grades Chicago Alpha Averages 13-5.6 14-6.0 108.0 81.4 35.1 87.5 Medians 13-3.0 14-6.5 107.0 81.5 36.5 79.0 *RUCH, G. M. : An Experiment with Forced Promotions, Educational Administration and Supervision, 6, 1920, pp. 71-73. [35 Study of Pupil Development "Examination of the figures given above leads to the eonclu- tion that the promotions could be defended except in the case of H. C. who did passing work, average grade of 76 (passing 75) although at a high expenditure of effort. In no other case did a pupil fall below the median grade of the school (median for school 82 ; boys, 80, girls 83) . The one pupil doing poor work would have been eliminated on the basis of the supplementary evidence fur- nished by the intelligence tests had these been possible at the same time that the pedagogical tests were made. ' ' This illustration, even in spite of the small number of cases involved, tempts one to conclude that the careful use of the results of a combination of mental, educational, and physical tests would offer a basis for forced or special promotions, where neces- sary, which is about as reliable as educational practice can be made empirically. In the foregoing example, much valuable supplemen- tary data is again overlooked in the direction of physical factors. To take a single case, that of the pupil not advanced because of physical factors, we have the following : Score Pupil— H. M., Boy— Attempts Eii C. A. 11-11 Courtis M. A. 14-0 Addition 8 8 I. Q. 117 Subtraction 12 12 Alpha Chicago Average Grade 83 38 83 Multiplication Division 5 6 5 6 Weight 72 Kelly Height Vital Capacity Vital Index 52.5 130 1.80 Gregory Grip — Eight Grip— Left 27 29 17.8 52.6 Examination of this data shows that the pupil rates quite satis- factorily in all of the tests of general intelligence, fairly high or very high in the standard educational tests, but that there is some evidence of physical immaturity, particularly in height. This is not serious in relation to the norms for eastern children, but is somewhat subnormal for the pupils of this school. Taking into account the fact of the low age of the pupil as well, it was decided not to permit him to make the advance in grade. Another case which shows the influence of physical factors is that of a boy who is one year accelerated but who has failed to maintain the high stand- ards of work evidenced by accelerated pupils. His measurements follow : 36] University of Oregon High School S. F.- -Boy. C. A. 12-10 Courtis Attempts Bights Score M. A. 11-8 Addition 6 2 I. Q. 91 Subtraction 6 6 Alpha 54 Multiplication 8 5 Chicago 361/2 Division 7 6 Average Grade 75 Weight 63 Kelly 22.3 Height 56 Vital Capacity Gregory 18.8 Vital Index Ay res Spelling 80.0 Grip — Eight 22.5 Ayres Writing 40.0 Grip — Left 22.0 Stone Reasoning 9 6 10.5 This boy is somewhat sub-normal in intelligence although not seriously enough to explain his poor school work. His previous school advantages have not been of the best but this also is not a serious defect. The real reason for his poor school work is to be found prominently involving the fact that his mentality will not warrant his continuance in a grade one year above the normal for his age and further that his physical development is retarded. The boy is almost the tj^pical picture of the cases of mal-nutrition as described in works on school hygiene. It is not accurate to speak of him as handicapped by ill health as he is rarely sick and is regular in attendance at school, but, casual inspection reveals that meta- bolism is carried on at a very low ebb as is evidenced by his apathetic manner. In all probability he will soon lose his advantage in school position and through failure drop back into his proper grade. It seems to be one of those cases where a teacher can con- scientiously advise the parents to allow the boy to be retained for one and a half or two years in grade eight before allowing him to enter high school. It is even possible that parents would be glad to pursue this plan if presented as outlined here with the facts before them and in a sympathetic manner, with some reservations, of course, in the m^atter of revealing intelligence scores. These case descriptions suggest at once the problems of age- grade distribution as a measure of school progress as dependent upon factors of native endowment. Acceleration in most schools in the past has not attracted a great deal of attention, due to the fact that our methods of promotion did not favor the production of accel- eration to any such degree as they did produce conditions of retard- ation. It has been shown, as might have been expected, that acceler- ation is closely related to superior ability. Our figures presented before point unmistakably to this conclusion. That the reverse would hold true of retardation has not been nearly so well accepted [37] study of Pupil Development because of the multiplicity of factors involved in causing pupils to fall behind grade. However, there is a growing suspicion that many of the conventional causes to which retardation has been attributed are after all the superficial causes which often conceal the real reasons. For this reason it was thought worth while to attempt to analyze our cases of retardation in the light of the experimental facts gathered. One of the striking facts is that of the relative ranks of the accelerated, at age, and retarded groups in the various types of tests. In general the most perfect gradations are found to be characteristic of the tests of general intelligence (see table 18), school marks (see table 20), and percentages of failures (tables 21 and 22) and is least characteristic of the stand- ard educational tests and the phj^sical measurements (tables 26, 28, and 30). In order to make certain of these comparisons directly portions of the foregoing tables are reproduced here. TABLE 44 Showing in summarized form the results of the tests of intelligence and school marks for the groups of accelerated, at age, and retarded pupils of the entire school. All medians. N. C.A. M.A. I.Q. Per cent Alpha Chicago Grades failing Accelerated 40 Normal 44 Retarded 35 13-2 13-7 15-9 15-2 15-0 14-1 117 105 90 104 85 79 43 83 7.3 35.5 82.5 9.1 31 78 31.4 TABLE 45 Test Grade FII Accelerated At age Betarded Grade VIII Accelerated At age Betarded Kelly Courtis (totals) Gregory Ayres Writing Ayres Spelling 20,5 17.0 34.3 40.0 60.0 21.1 21.0 20.0 40.3 81.7 20.3 12.0 22.1 55.0 75.0 29.4 25,5 44.2 45,0 96.0 32.0 19.1 29.0 20.5 32.0 28.2 65.0 55.0 92.5 86.7 In examining the foregoing summaries one cannot fail but bo impressed with the weight of native ability in the determination of school standing. One thing is certain, viz., that mental deficiency is a far more powerful factor in the cause of retardation than has been commonly supposed notwithstanding several authorities to tV.e contrary who asserted that this is a minor if not negligible influence. Concretely to the point is the matter of late entrance. Why should delayed entrance to the public schools result in a lessened acquisi- tion of classroom knowledge, all other things being equal, in com- [38] University of Oregon High School parison with the pupil who begins his school life on schedule time 1 He must have traveled the same road through the school, studied under the same instructors, followed the same course of study, etc. It must be true that such groups of late entrants within the ranks of the retarded cannot account for the generally poor showing of the latter. In the case of ill health, moving away, and similar causes, our contentions are not meant to apply. These are admittedly real causes for pupils falling behind grade. However, as we shall show below, there is a strong tendency in the case of loss of time through changing schools, for the factor of intelligence to enter once more. The floating population repre- sents a class very largely recruited from the ranks of those who are to be described as economic failures to a greater or less degree. Here, even though such frequent changes are real handicaps to successful school work, the underlying cause reduces often to the inability of such homes to maintain a satisfactory stand- ing in the economic competition of the industrial world. The eco- nomic instability is but the superficial cause. In order to check upon the weight to be assigned to the factor of native ability in individual cases of retardation as well as in the mass as shown by the tables of medians above, the full school history of all the cases of retardation remaining in the school was obtained. Of the thirty- five cases of retardation in the school during the first semester of the school year 1919-1920, thirty-one of these pupils continued throughout the year. In order to examine the attributed causes for these pupils having fallen behind grade, the detailed school history of each case was secured by the combined use of information blanks, personal interviews, and conferences with parents. The school progress of each pupil was then charted out from the time the pupil entered the first grade to the present time. That a few errors inev- itably have crept in matters very little since only the attributed causes are to be considered and these are purposely to be taken at face value. It is not believed that the attributed causes are gener- erally the true causes, on the other hand, it was quite often evident that the contrary was true. The real intention is to subject the tradi- tional attributed causes to a further analysis in the effort to show- that such causes are but superficially operative and that in many cases deeper-seated forces have operated to cause the pupil to make slow progress through the schools. Table 46 shows in a summarized form the inter-relations between the attributed causes and such factors as the intelligence quotients and the number of schools at- [39] study of Pupil Development tended. Notwithstanding many exceptions, the number of schools attended should in the mass prove to be a rough measure of the economic instability and restlessness of the home from which the pupils have come. That some relation between the factor of intelli- gence and the floating home exists has been pointed out by numer- ous investigators. TABLE 46 Giving the median I. Q. 's, absolute numbers, percentages, and numbers of schools attended for the groups as classified upon the basis of attributed causes for the 35 retarded pupils. Aver. No, Per Median of schools Attributed Cause N. cent I.Q. attended. I. Single Causes: Absence due to changing schools or being put back 6 19.3 89.0 6.17 Absence due to ill health or accidents 3 9.7 90.0 3.67 Service in army 1 3.2 101.0 3.00 Late entrance 4 12.9 100.5 3.25 Inability to learn 5 16.1 86.0 4.40 Lack of application 7 22.6 90.0 4.00 II. Multiple Causes: Absence due to changing schools and ill health combined 1 3.2 90.0 6.00 HI health and lack of application combined 2 6.5 99.5 5.00 Late entrance and poor health combined 2 6.5 74.5 4.50 III. Left School Before Data Was Secured : Transferred to another school 1 62.0 To go to work 2 80.0 Moved away 1 35 Examination of the foregoing table shows the suggestive fact that no matter what the attributed cause of retardation may be, that the median intelligence of no group rises above mediocrity and in the vast majority of the classes it falls decidedly below the nor- mal. The highest levels of intelligence accompany such factors as late entrance, army service delays, and ill health coupled with other causes. It is quite evident that none of these groups would logically be expected, per se, to be characterized by low grade ability in the same sense that lack of application or inability to learn would be. This data, although slight in amount, would tempt one to adopt a conclusion that is diametrically opposed to the statement made ny some students of retardation that defective mentality is a slight factor in slow progress through the schools. Rather would poor native endowment appear to underlie most of the traditional causes [40] University of Oregon High School for these conditions. Terman* quotes Dr. Gulick as taking the posi- tion "that relatively few children are so defective as to prevent success in school or life," but accepts the results of Dickson whose ^dings are quite in accord with the facts presented above. One thing seems to be certain, at any rate, and that is that the whole subject of the causes of retardation must shortly be re-opened in the light of the new knowledge in the field of individual differences and tests of endowment. This relation between school progress as rated in terms of accel- eration and retardation and the factor of native ability or general intelligence suggests the problem of comparing the scores made iji the standard educational tests with ability by the method of correla- tion. The exact degree of relation between school acquisition and real ability has been a much disputed question in the past, and only -recently has much experimental evidence been available for pur- poses of discussion of such points. In table 47 presented below are given the Pearson coefficients between the I. Q. and the scores in the pedagogical tests, grade by grade. A few statements concern- ing the method of handling the data for these correlations are needed. The method of averaging the coefficients for the three grades (in part of the cases, two grades) was that of a "weighted" average in which the exact size of N was allowed to influence the average. This is obviously a truer measure than a direct average of the two or three coefficients. It may be objected that even such averages are not very reliable. This is admitted but, on the other hand, it is thought that such weighted averages are in most cases lower than the true correlations. For example, the correlation for all pupils of all three grades between the I. Q. and the average grades earned is 0.533 plus or minus .044 but the weighted average correlation is 0.480. The use of the weighted averages of the coeffi- cients makes it possible to avoid certain errors which would arise in comparing scores in a given test for several different grades, viz., the I. Q. varies about 100 in each grade while the test scores vary about different bases in each grade, usually a series of increasing values as we pass from a grade to the next higher, etc. This fault, ol course, does not apply in the case of the average grades since the grading system has a constant base at all grades. Since the I. Q. is an index and not a score, it could only be used for correlations within a grade for comparison with scores but might be used for all grades for comparisons with other indices (the school marks are ♦Intelligence of School Children, p. 119. [41 Study of Pupil Development obviously indices). The mental age, in contrast, is a true score and might be used but it is open to the objection that the same mental age score might stand for pupils representing all three grades and hence introduce an error by virtue of the fact of unequal training effects in the different grades. This effect would most likely be most serious in subjects receiving direct training in the grades under discussion, e. g., language, and least serious in those subjects where the direct training has ceased several grades below the ones here concerned, e. g., in the Courtis tests. These considerations form the argument for the use of the weighted averages of the Pearson coefficients. In the Stone tests only the reasoning scores are concerned. In the Courtis tests the scores for the four operations are combined. This is a purely arbitrary procedure, but since a point gained in addition is not so very unequal to one for subtraction or multiplica- tion, the errors may tend to balance one another to a large extent. At any rate, for our purposes, it was desirable to have a single value to stand for the ability in the fundamentals of arithmetic processes. Table 48 shows the weighted average correlations arranged in order of magnitude. TABLE 47 Showing the Pearson coefficients of correlation for the intelligence quotients and the scores made in the standard educational tests by grades. Grade VII Grade VIII Grade IX Name of Test r P.E. r P.E. r P.E. ,Stone (Reasoning) 0.715 .055 0.590 .066 Gregory 0.559 .082 0.661 .057 Average Grades 0.396 .099 0.535 0.65 0.471 .093 Courtis (Sums) 0.342 .103 0.306 .087 Kansas S. E. 0.298 .109 0.702 .052 Ayres Spelling 0.080 .116 0.263 .094 0.298 .114 Ayres Writing —0.350 .102 —0.034 .101 —0.159 .120 Douglass Algebra 0.397 .481 TABLE 48 Showing the weighted averages of the coefficients for the I. Q. and the various tests arranged in order of magnitude. Test Weighted Average Coefficient Stone Reasoning scores 0.65 Gregory Language Test 0.62 Kansas Silent Reading 0.53 Average Grades 0.48 Courtis ( Sums) 0.32 Ayres Spelling 0.22 Ayres Writing — 0.17 [42] University of Oregon High School The correlation between the grades earned and the mental age ratings is 0.42, P. E., 0.76, which is slightly below the coefficient which Proctor reports for the 111 ninth grade pupils which he stiidied (r equals 0.45). The number of cases used for our correla- tion was 117. If the pupils of the school are classified upon the basis of the I Q. as before, into groups of superior, average and inferior, the correlations with the mental age for school marks are : Median Median Group N- r M. A. Grade Superior 50 0.23 16-3 84.1 Average 48 0.29 14-4 80.4 Inferior 18 0.18 11-11 77.5 It seems to be true that pupils who are to be classified as of average ability are somewhat more likely to realize results in school work according to their ability than are either very able or very dull pupils. The medians for mental age and marks are quite uniformly graded among the three groups and suggests that such data can throw considerable light upon a problem which is of prime impor- tance in school practice, viz., the realization of the maximum efforts of each pupil and of the establishment of just standards of attain- ment. With the use of a grading system like the one in use in the University High School where the assignment of school marks must follow approximately the distribution of the normal curve, it is possible to equalize the distribution of grades and to attain a standard which is relatively just, i. e., the pupils can be graded with respect to each other. This plan has generally proved more work- able than attempts at grading upon any absolute basis. The rela- tive grading plan does, however, require that the standards be "set" according to the abilities of the group of pupils in such a way as not to be unfair to any particular unit of the pupils, e. g., to discriminate against the dull group, or to prevent differentiation at the upper levels in advent of low standards. In order to attempt to evaluate the standards used for grading in the University High School, the actual distribution of the assigned grades will be compared with the best-fitting normal curve for this actual distribution. The table below shows the actual number of each letter grade given, the theoretical number assigned to the best-fitting curve, and the approximate numbers for the [43] study of Pupil Development normal curve. In using the letter grades, the mid-points were taken as before as 97.5, 92.5, 87.5, etc. TABLE 49 Approximate Letter Theoretical theoretical Actual Grade numbers ntimhers numbers A+ 1.56 2 1 A 9.73 10 4 B+ 29.55 30 30 B 43.75 44 51 C 31.40 31 24 D 11.05 11 13 E 1.89 2 6 Average, 82.29. Sigma, 5.89. N=129. It will be noted from these figures that the actual distribution of grades is skewed slightly from the normal toward the lower end of the scale. This is in part due to the fact that the average grades are somewhat lower than the true grades as explained in a preced- ing section, and partly due to the super-normal numbers of very low grade pupils in the school (see tables of distributions for the intelligence quotients). On the other hand, the very large numbers of exceptionally high grade pupils which have been described as characteristic of this school, fail to reveal their presence in the upper ranges of the grading scale. The skewness of the curve would seem to mean either one or both of two alternatives : 1. That the standards are luiiformly too severe in this school, or 2. That the high grade pupils are not doing their maximum quality of work. There is some evidence to be had for both views. In the first place, 13 C and 6 D grades were given — a total of 19, whereas the theoretical indicates but 13 failures (D's and E's). The numbers of failures is therefore somewhat excessive. On the other hand, but five pupils receive A or above when 12 should score that much according to the normal distribution figures. This, in view of the marked skewness of the curve of the distribution of the I. Q.'s at the upper ranges, would indicate that the work of the most intelli- gent pupils does not approach the maximum attainment to as per- fect a degree as for other groups. The truth of this conclusion again must be tempered in the light of the fact that children of superior ability tend to be accelerated and are thus subjected to [44 1 University of Oregon High School greater competition than normal children. All in all, the evidence does not seem to point to the existence of any considerable injustice to any group arising from the standards of scholarship in the school, but it does show that with more attention to the assignment of grades there could be maintained a somewhat closer approxima- tion to the theoretical curve, particularly in the extreme ranges. As to the possibility that the most able pupils do not realize as fully their highest possible attainment as do the normal and inferior groups, the evidence follows : TABLE 50 Giving the median scores for the Superior (I. Q. 's 111 or above), Average (I. Q. 's 90-110), and Inferior (I. Q.'s 89 or less) groups for all grades. Sexes combined. Average Group No. C. A. M. A. I. Q. Alpha Chicago Grades Superior 50 13-1 16-2 121.0 108.0 45.0 83.0 Average 50 14-4 14-3 101.0 81.0 36.0 81.0 Inferior 20 15-8 11-10 84.5 65.0 25.5 77.0 It will be noted that although the average pupil earns grades four points above the inferior, the superior pupil earns but two points more than the average. In practically all of the scores except the grades the medians for the superior group are higher by a larger margin from the average than the latter are in turn from the inferior, but at the same time it must be remembered that the superior pupils are actually younger than the others and are likely to be accelerated and hence subjected to more intense competition. The evidence is therefore not very conclusive. In conclusion it should be pointed out that many other lines of investigation, conclusions, applications, and the like, have been suggested which, for reasons of the limitations of space, must be neglected. Specifically there are certain questions involving the diagnostic values of tests of intelligence or of pedagogical accom- plishment which are deserving of attention. One such case was pointed out in passing in the matter of the discovery of the reasons for the poor showing of the three arithmetic sections in the funda- mentals of arithmetic. Here it was possible to check up the testi- mony of the teachers concerned, with the more objective measures of experimental education. That the teachers were agreed upon the position that the fault was to be traced to poor instruction in the lower grades of the elementary- school suggested the plan of check- ing the abilities in the four fundamental processes as measured by [45] Study of Pupil Development the Courtis tests against the more intellectual processes of reasoning and inference brought into play in the tests devised by Stone. In case of the latter, as well as in the former, the accuracy scores revealed again the early weakness of formal instruction in this subject, but at the same time, indicated a tendency toward improve- ment under the force of later instruction and also that there was no sub-normal accomplishment in the types of problems characteristic of the higher parts of the elementary school curriculum. Actual tracing back of the difficulty through the lower grades in consulta- tion with the Superintendent of the Eugene Schools revealed cor- roborative evidence that these pupils had received the major portion of their instruction in the fundamentals of arithmetic at the hands of an unskilled teacher who was later dismissed from the system. This evidence establishes almost beyond doubt the correctness of the inferences drawn from examination and comparison of the scores from the two standard arithmetic tests. Other questions of the diagnostic values of the various types of tests might well be considered if space were available, since there is every indication that this phase of school measurement is but be- ginning to attract the attention of workers in the elementary field. Diagnosis will in all probability be a prime characteristic of the newer tests and the newer methods of experimental education. If any new light has been shed upon the application of experimental methods in the course of this survej^ of a single school, the efforts expended are well repaid. SUMMARY Briefly stated the main conclusions presented in this discussion are: 1. In view of the value of the results of intelligence testing from the administrative angle, the amount of time required for tests cannot reasonably be held to be prohibitive. 2. Group tests will not ordinarily be found satisfactory for individual diagnosis, but will usually suffice for the purposes of mass measurements. 3. Teachers' estimates of intelligence are not likely to prove substitutes for individual testing, even if the reliability of such estimates might be very high, because the data on those pupils mis- judged by their teachers is the very data which are most needed. [46] University of Oregon High School 4. The pupils of the University High School were found to be a composite group composed of two distinct elements ; first, pupils from University faculty homes and other homes of professional people, and secondly, pupils representative of homes of low economic level. These factors must necessarily be reckoned with in the inter- pretation of certain of the findings. 5. Group test "Army Alpha" was found to have a somewhat greater reliability than the "Chicago Group Intelligence Test" in comparison with the Binet tests. 6. Retardation as a school problem must shortly be re-studied in the light of the newer facts of individual differences brought forward by the use of intelligence tests and other objective measures. 7. In the majority of cases, pupils classified as retarded by the usual standards are really accelerated upon the basis of mental age, and the reverse is true for accelerated pupils. 8. Accelerated pupils, although actually younger, are able to more than maintain their position in competition with children who are "at age" as is shown by school marks, but retarded pupils do not succeed in holding their own in their class in spite of their greater age. 9. Retarded pupils furnish about three-fifths of the total fail- ures of the school. 10. The pupils of the University High School show satisfactory abilities in all school subjects tested except arithmetic and writing. In the case of the former the reason was proved to lie in poor instruction in the lower grades. 11. Boys are slightly inferior to girls as measured by the grades received in school. The range of grades is somewhat larger in the case of the girls. 12. The results of the Stone tests seem to indicate that the diffi- culties in arithmetic are to be diagnosed as inability to handle the fundamental processes in contrast with the ability to reason. 13. The physical measurements show that the Oregon pupils are both taller and heavier than eastern children. The only pub- lished norms which approximate the results for the western children are those of Baldwin. 14. The most marked superiority of the University High School pupils in the physical traits studied was found to be the much better development of lung capacity and the vital index. [ 47 ] Study of Pupil Development 15. The new method of correlating physical and mental traits advocated here seemed to indicate the presence of correlation be- tween the vital index and intelligence. 16. Strength of grip did not seem to show relation with intelli- gence, at least not in the case of girls. 17. The possibility of the use of intelligence tests and tests of pedagogical accomplishment as the basis for promotions, grading, and similar questions of school progress, is rapidly becoming evi- dent as a method of surmounting the lock-step evil in education. 18. Forced promotions, where unavoidable, can be made with a high degree of reliability by the use of objective measures. 19. Correlations of abilities in school subjects with general intelligence show the highest coefficients in the language and rea- soning tests and the lowest in spelling and writing. School marks and arithmetical ability are intermediate. 20. The grading system of the University High School is prob- ably too severely enforced at the upper ranges as is shown by super- imposing the best-fitting normal curve on the actual distribution. There is also some evidence that the superior child fails to approach maximal efficiency as closely as do less able pupils. [48] LIBRARY OF CONGRESS e 019 842 631 3 • HoUinger Corp. pH8.5 f HoUinger Corp. pH8.5