The Fifteenth Yearbook OF THE NATIONAL SOCIETY FOR THE STUDY OF EDUCATION Part I STANDARDS AND TESTS FOR THE MEASUREMENT OF THE EFFICIENCY OF SCHOOLS AND SCHOOL SYSTEMS THIS PART OF THE YEARBOOK WILL BE DISCUSSED AT THE DETROIT MEETING OF THE NATIONAL SOCIETY, MONDAY, FEBRUARY 21, 1916, 8:00 P.M. IT WILL ALSO BE DISCUSSED AT THE DETROIT MEETING OF THE NATIONAL COUNCIL OF EDUCATION THE UNIVERSITY OF CHICAGO PRESS CHICAGO, ILLINOIS THE UNIVERSITY OF CHICAGO PRESS CHICAGO, ILLINOIS THE CAMBRIDGE UNIVERSITY PRESS LONDON AND EDINBURGH THE MARUZEN-KABUSHIKI-KAISHA TOKYO, OSAKA, KYOTO KARL W. HIERSEMANN LEIPZIG THE BAKER & TAYLOR COMPANY NEW YOSK The Fifteenth Yearbook OF THE NATIONAL SOCIETY FOR THE STUDY OF EDUCATION Part I Standards and Tests for the Measurement of the Efficiency of Schools and School Systems BY G. D. Strayer, Bird T. Baldwin, B. R. Buckingham, M. R, Trabue F. W. Ballou, D. C. Bliss, H. G. Childs, S. A. Courtis E. P. Cubberley, Charles H. Judd, George Melcher E. E. Oberholtzer, J. B. Sears, Daniel Starch, G. M. Whipple Edited by Guy Montrose Whipple, Secretary THIS PART OF THE YEARBOOK WILL BE DISCUSSED AT THE DETROIT MEETING OF THE NATIONAL SOCIETY, MONDAY, FEBRUARY 21, 1916, 8:00 P.M. IT WILL ALSO BE DISCUSSED AT THE DETROIT MEETING OF THE NATIONAL COUNCIL OF EDUCATION THE university OF CHICAGO PRESS CHICAGO, ILLINOIS ^-^s ,11^ Copyright 1916 Bv Guy M. Whipple SECRETARY OF THE SOCIETY All Rights Reserved Published February 1916 Composed and Printed By The University of Chicago Press Chicago, Illinois, U.S.A. OFFICERS OF THE NATIONAL SOCIETY FOR THE STUDY OF EDUCATION President Randall J. Condon Superintendent of Schools, Cincinnati, Ohio Vice-President J. Carleton Bell University of Texas, Austin, Texas Secretary- Treasurer Guy M. Whipple University of Illinois, Urbana, Illinois Executive Committee (The year indicates date of expiration of term) William C. Bagley (1916) University of Illinois, Urbana, Illinois George D. Strayer (1917) Columbia University, New York City Harry B. Wilson (1918) Superintendent of Schools, Topeka, Kansas DwaGHT B. Waldo (1919) State Normal School, Kalamazoo, Michigan Board of Trustees Charles A. McMurry (1916) State Normal School, De Kalb, Illinois Lotus D. Coffman (1917) University of Minnesota, MinneapoUs, Minnesota S. Chester Parker (1918) University of Chicago, Chicago, Illinois 31«^n9 This Yearbook is the 191 6 report of the Committee of the National Council of Education of the National Education Association on Standards and Tests of Efficiency. MEMBERS OF THE COMMITTEE George D. Strayer, Chairman Professor of Educational Administration, Teachers College, Columbia University Bird T. Baldwin Professor of Psychology and Education, Swarthmore College Adelaide S. Baylor Assistant State Superintendent of Schools, Indiana Ben Blewett Superintendent of Schools, St. Louis, Missoiuri Ellwood p. Cubberley Professor of Education, Leland Stanford Junior University Edward C. Elliott Chancellor of the University of Montana Paul Hantjs Professor of Education, Harvard University Charles H. Judd Director of the School of Education, University of Chicago Calvin N. Kendall Commissioner of Education for the State of New Jersey Charles H. Keyes President, Skidmore School of Arts, Saratoga Springs, New York William H. Maxwell Superintendent of Schools, New York City, New York John H. Phillips Superintendent of Schools, Birmingham, Alabama Frank E. Spaulding Superintendent of Schools, Minneapolis, Minnesota Edward L. Thorndike Professor of Educational Psychology, Teachers College, Columbia University James H. Van Sickle Superintendent of Schools, Springfield, Massachusetts TABLE OF CONTENTS PAGE Editor's Preface 7 ixtroduction 9 George Drayton Strayer, Chairman of the Committee PART I. EDUCATIONAL SCALES AND UNITS OF MEASUREMENT CHAP. I. A Measuring Scale for Physical Growth and Physio- logical Age II Bird T. Baldwin, Professor of Psychology and Education, Swarth- more College II. Notes on the Derivation of Scales in School Subjects: With Special Application to Arithmetic .... 23 B. R. Buckingham, Chief Statistician, Department of Education, New York City III. Score Card for City School Buildings 41 George Drayton Strayer, Professor of Educational Administra- tion, Teachers College, Columbia University IV. Completion Tests for Public-School Use . . . . 52 M. R. Trabue, Instructor in Educational Administration, Teachers College, Colimibia University PART n. APPLICATION OF SCALES AND UNITS OF MEASUREMENT IN EDUCATIONAL SUPERVISION ANT) ADMINTSTRATION V. Work of the Department of Educational Investigation and Measurement, Boston, Massachusetts . . . . 61 Frank W. Ballou, Department of Educational Investigation and Measurement, Boston, Massachusetts VI. The Application of Standard Measurements to School Administration 69 D. C. Bliss, Superintendent of Schools, Montclair, New Jersey VII. A Half-Year's Progress m the Achievement of One School System A. The Progress as Measured by the Thorntjike Visual Vocabulary Test 79 B . The Progress as Measured by the Courtis Tests, Series B 83 H. G. Chuds, Associate Professor of Education, Indiana University 5 6 THE FIFTEENTH YEARBOOK CH PAGE VIII. Courtis Tests in Arithmetic: Value to Superintendents AND Teachers 91 S. A, Courtis, Supervisor of Educational Research, Detroit, Michigan IX. Use of Standard Tests at Salt Lake City, Utah . . 107 Ellwood p. CtTBBERLEY, Professor of Education, Leland Stanford Junior University X. Reading m Charles H. Jodd, Director of the School of Education, University of Chicago XI. Studies by the Bureau or Research and Efficiency of Kansas City, Missouri 120 George Melcher, Director of the Bureau of Research and Effi- ciency, Kansas City, Missouri Xn. The Effects of Efficiency Tests in Reading on a City School System 138 E. E. Oberholtzer, Superintendent of Schools, Tulsa, Oklahoma XIII. Investigation of Spelling in the Schools of Oaklaisod, California 141 J. B. Sears, Leland Stanford Junior University XIV. Standard Tests as Aids in the Classification and Promotion OF Pupils 143 Daniel Starch, Assistant Professor of Psychology and Education, University of Wisconsin XV. The Use of Mental Tests in the School 149 Guy Montrose Whipple, Professor of Education, University of niinois Constitution of the National Society for the Study of Education 161 MmuTES of the Cincinnati Meeting of the National Society for THE Study of Education 163 Financial Report of the Secretary -Treasurer of the National Society for the Study of Education 165 List of Active and Honorary Members of the National Society for the Study of Education 167 Application for Membership in the National Society for the Study of Education i73 EDITOR'S PREFACE It will be recalled that the National Society for the Study of Educa- tion one year ago published as its Fourteenth Yearbook, Part I, the 191 5 report of the Committee of the Department of Superintendence of the National Education Association on Economy of Time in Education, Superintendent H. B, Wilson, chairman. It was anticipated at that time that the policy of this Society in publishing Yearbooks before the meeting at which they were to be discussed would measurably augment interest in' the report and increase the effectiveness of its discussion. This anticipation was so fully realized that, now that the further report of the Committee on Economy of Time has been unavoidably delayed, it is a matter of satisfaction to be able, nevertheless, to continue our poUcy by presenting to our members, as Part I of the Fifteenth Yearbook, the report of another important committee, the Committee of the National Council of Education of the National Education Association on Standards and Tests of EflSciency, Professor G. D. Strayer, chairman. The perusal of the Table of Contents will suffice to show that this Yearbook maintains the high standards set by its predecessors in the selection of subject-matter that bears directly upon educational prob- lems of recognized importance and in the selection of contributors that guarantee a presentation worthy of careful consideration. G. M. W. INTRODUCTION In the pages which follow, the Committee on Standards and Tests of the National Council of Education presents a series of reports pre- pared by its own members and by others who have, upon invitation of the committee, been willing to contribute results of their study or investigation. The committee feels that these reports furnish most satisfactory evidence of the progress which has been made in the field of educational measurement during the past few years. The papers of the several contributors may, in general, be classified as contributing (i) to the derivation of scales or units of measurement, and (2) to the application of such units or scales of measurement as have been derived in the field of educational administration or supervision. In printing, the papers have been grouped as indicated above. It seems unnecessary with this report in hand to argue in favor of the use of the more precise methods of measurement which have recently been developed. Those interested in the improvement of our schools have always attempted by some method or other to estimate the effi- ciency of individual schools or school systems. Often this measurement has been based upon an opinion not acceptable to another student or investigator, and not frequently upon attempts to measure which have been extremely crude. The examination system as we have commonly had it, not only has been responsible for judgments with respect to the efficiency of schools, or of school systems, but has also determined in very considerable measure in most school systems the progress of children. In the studies which are reported here, more precise measures give to the student of education a better basis in knowledge of conditions upon which to base his criticism or to develop his improved method of instruction or administration. The measurement of results of any sort, whether of the achievements in school subjects, of the cost of any unit or fimction, or of the rate of progress, and the like, furnishes primarily a knowledge of the situation which makes clear the problems involved, and which may suggest a method of experiment that looks toward the improvement desired. Regardless of the development which may be made in the field of measurement, we shall always have to deal with the problem of aim in 9 lO THE FIFTEENTH YEARBOOK education. Those who have done most to develop precision in measure- ment, or who have profited most by using the units or scales, would be the last to deny the worth of that thinking and discussion which leads to a determination of the ends to be realized in our schools. When one has defined purposes or ends to be achieved, efficiency, in the light of these aims, can be determined only as we are able to measure the degree to which improvement or growth has taken place. Indeed, education may be best defined in terms of changes which are brought about in the individuals subjected to the process. If our aims mean anything to those who teach, they are or are not satisfied, in proportion as measur- able changes are brought about in pupils. The more precisely we are able to measure the development which is brought about by virtue of our school work, the more certain we may be that we are realizing the aims which we have set up, and it is only by means of some sort of measurement that we may even claim to have made progress in the accomplishment of those ends which we profess to seek. George Drayton Strayer, Chairman PART I EDUCATIONAL SCALES AND UNITS OF MEASUREMENT CHAPTER I A MEASURING SCALE FOR PHYSICAL GROWTH AND PHYSIOLOGICAL AGE BIRD T. BALDWIN Professor of Psychology and Education, Swarthmore College A comparative study of the results of four hundred investigations on over one million individuals shows that the height, weight, and lung capacity of children vary according to nationality, heredity, general social status, urban and rural conditions, geographical distribution, season of the year, normal and abnormal mentality, health, sex, initial size, and stages of physiological maturity. Among the marked modi- fying factors which have, to a considerable extent, made data incom- parable are to be found the manner in which measurements have been taken and recorded; the personal equation of trained and untrained examiners; the varying degrees of accuracy of measuring apparatus; the age of the child, as based on exact birthday, last birthday, or next birthday, with or without months, weeks, and days being taken into consideration; and the measurement of children with or without clothing. This paper aims to formulate tangible norms which may he used by physical directors and teachers as standards for comparison with all types and races of children between the ages of five and one-half and eighteen years. The norms are based on the best available data in this country from the Horace Mann School, the Francis W. Parker School, and the University Elementary and High schools of the University of Chicago, where a limited number of American children were measured consecutively for several years by trained anthropometrists using standardized apparatus of minute units, with the age recorded exactly in days. Details and 12 THE FIFTEENTH YEARBOOK source material have been published elsewhere/ but the material and charts included in this article are new and supplementary to the previous larger study. GENERAL PRINCIPLES OF GROWTH These norms, in accordance with facts previously published by the writer, conform with the conclusions that: Boys are taller, heavier, and have better lung capacity than girls except during the early adolescent period, when the converse of this statement is true. The widest ranges of individual differences and the largest increments of growth are during adolescence. Individual growth curves previously published show that in the course of growth short children do not become tall, neither do tall children become short, under normal conditions; each child holds his or her relative position in reference to a given median or norm through- out the school age — that is, in a well-developed child weight, height, and lung capacity are relatively proportionate to each other. If a child's weight is divided by its height, the weight-height coefficient is found, which is approximately the same for a well-developed large child or a similarly well-developed small child; the same relationship holds true for the breathing capacity and height or the so-called vital-height coeffi- cient. (See score card for norms.) STAGES OF PHYSIOLOGICAL MATURATION OR PHYSIOLOGICAL AGE Growth is a continuous process, but some periods in the life of a child are marked by more acceleration than others. The period from six to seven and a half is a period of rather rapid growth in height. Boys and girls above median or average height between the ages of six and eighteen grow in stature and physiological maturity in advance of those below the median or average height. There is a shifting of the acceler- ated period according to the individual's relative height, weight, and lung capacity. " Our study of thirty thousand measurements in height, weight, and lung capacity reveals correlations in growth for boys and girls above the median or average height different from those below. The rhythms and fluctuations of growth in height for the children above the median show that these boys and girls are physiologically older than those ^ B. T. Baldwin, "Physical Growth and School Progress," U.S. Bureau of Educa- tion, Bulletin No. lo, 1914. Whole No. 581. 215 pp. PHYSICAL GROWTH AND PHYSIOLOGICAL AGE 13 below the median, since their periods of acceleration and arrest begin earlier and end earHer. Short children continue their growth longer than tall children. nn " 1 ! ! 1 1 I 1 n " i -n 1 ' i t i 1 i 1 1 1 1 1 1 1 ! ill! — f^ -0- — 0 iNpIV 10 Ml GlWUTH CURVCS 1 \Jr^\^ 1 ;IN H;ei6HT ! M VI ! i l-J—M — ^' £6 wit-h ,shisqes of PnySiolcqifcJ;, /' 1 ^'i i 1 H o( ur jhon ilndici ihed / .--Ut-^i — ° fblBiLS) / ^ ^---^ -.i-^ 1/ /iyf^Vi- >^ 1 1 1 L-i / > ' [> .///-^ ^ ^^J^ ■' ^ / < /^ ^i'-r y / / i >^ X ^ r /J / a } /I t/ i;^ . / / / y 5 / ' / ^ ^ 1 ^ •/ / /y/ /\ y ^R B 1 { / t p y^y/ /^ -; 1 i / / /I //!^ / 54 ir i yl/ / uKj ^ /f f /I = // ^ / ^ 4/ / / x^ 52 _l2 / _d X /f/^ ^ 1 / P 4 ■— c c o < •: w ■« "2 in \ \ i l' V V *«., X 1^ \ v^ ^ "X 1 /f Q> U 5 X ^ y^ <^ B s > 1 S ^» « K> y' 1 .,/ 1 • • 1 / ^ v^ CM * ,/ ' \ \ 1 ^ s ■ i > \ \ a 5 / /a <•> •a N # ■* 7 pj / f.UO 3jaf£ QO J 8 R o PHYSICAL GROWTH AND PHYSIOLOGICAL AGE 17 healthy children of accelerated physiological development be encouraged to proceed through school as rapidly as possible within the limits of thoroughness, and that the small, light children of retarded physiological APPEARANCE OF PUBESCENT CIL\NGES IN 1,241 GIRLS Age No. Pre- pubescent Percentage No. Pubescent Percentage No. Post- pubescent Percentage 6| 4 12 16 6 26 22 37 26 45 27 41 18 39 17 10 10 3 lOO.O 100. 0 100. 0 100. 0 lOO.O lOO.O lOO.O 100. 0 93-75 100. 0 78.84 62.06 58.20 39-53 15-15 15-38 4.83 7 7| 8 8h 0 9| 10 io| 3 6.25 II iii 10 II 16 15 25 25 II 8 5 3 2 19.23 37-93 23.88 34.88 37-87 38.46 17 -74 14-54 7.81 6.12 3-17 I 1.92 12 12* 12 II 31 30 48 47 58 45 61 43 43 32 33 9 25 15 18 4 14 165 745 17.91 25-58 46.96 46.15 77.42 85-45 90.62 91.83 96.83 100 0 J'Z I3| 14 141 IC ici I I 1-55 2.04 16 i6| 17 I7| 18 i8| IQ 100 0 4 20 20i 21 2l| 22+ I (22 71*5.) 362 .60 99-4 Total 134 Grand total. . . . 1,241 development be kept below or in the normal grade, doing supplementary work, since these short, light pupils are immature in mental develop- ment, although in many cases precocious in degree of brightness. It also follows from this study that rapid, healthy growth favors good mental THE FIFTEENTH YEARBOOK r — \\\ ~ [■"■ "" T"" ^ T"" "^ "~ ^ E c 1 |\ ^• V H^ ■\ c ^ 1 1 ^ ^ V X, 1 1^ %- N V \ 1 XI \ -s V, N. ■^ 1 K 1 § s \, 3 \ H *S >- s s ^ \ a A . W ^ \ \ V c \ y s ^ ►s s If \ A V K \ . S.^ < i Vv I i 1 \ \ 1 1 I V 1 ^~ \ V V ^ 1 Jiq u \ s ^ gS§§iSS§SiSS§£ ■ pu no 4 \ §2 §~ a2SS3{SSt5SSSS5«| 'I 1 1 \ V s V N ^ N -C S V V, 1 s v^ \ X2 -^a H X V ^ "- S ■V \ ^ ^ V \ E ^ J 1 \ a ■a u\ i \ . ^ 3 :^ 3 y. \ s ^^ s y. \ ^ ^r % \ d 1 2,S 3S \ V, \ N .1^= s \ CO a 5-^ ^ S^ _^ u- <■ > ' ~ _ ^ nj PHYSICAL GROWTH AND PHYSIOLOGICAL AGE 19 " - n "" "" ~ ■" ^ ~ k \ JO i . \ « \\ \ J N V \ ^ £ t: ^ ♦^ V >^ 1 i^ 2^ Si s ly-i I t s yv \ 2 4^ \ c s \ J \ o- s ~, X >^ '^ ! ' s N \ U i , r ■»», >, s. V c i s \ ^ V V V a \ 1 s \ «/• A S Si V -a A K \ ^ \ \ h ^ \ 1 V \ 1 0 ^-P ^ hsi ! f' |Ji 1 'M 0 ' \ ggggsii? ssssga \j\tn P<1 1 ^ ^ §3i~ £2|Ss{gS3gSSS4 5?| -J n 1 ! ! ( ' ~ L - \ ^ I b V \ 2 \ k 1* 'V ^ \ I . > \ \ k. T i ^ V, ^ s • •c •4 \ \ i ~ -c ~ s 7 £ 2 \ \ ^— G 5^ 3 ^ s- ^ ' ^ 1 •^ s \ e s a X e s. \ VI a: u 0 s \ -a . m a s ^ ~ < < t- ~\ V -^ \— ■~ r I M s \ b. ^ V ■^ Z V \ -^ s -^ s ~ ■~ \ s _ — L ^ ^ O] "Si ^ ^ o .5 Worn So 5 ^ 20 THE FIFTEENTH YEARBOOK O H w :^ o > Q ^1 J2 O o "a o c 55 Q •suuon inapgaoD iqSpq-jirjiA Diij3m=St9oo'X snijon jnaro -rgsoD iqSraq-prjiA qsyi^ng aqx 'suijou jnapcgaob iqSiaq-^qSiaAv Du?3ai = 6ii- X siujou '^uapcgaoD 'jqSiaq-iqSiaii qsjiSng 'aqx 'sjajij m xip^dBD 3citq;B3jq pn'B 'sm'Bj3oi!T( u; jqSp.si 'sjajatnpnaD ui jqSpq asn tuajsXs oujaui aqj nj •;u3ptg30D iqapq-iqS!3M= £6 • = 'u! ft-^'sqj i> 'sj-eaX S joj; 'aidurexa joj ' -^naprgaoD 5qSpq-p35iA=5q3iaq-i-Xii3BdB3 amqi^aig •^usptgaoD iqS!3q-iq8pA\=^iq3i3q^5q3p^Y St Bprauoj aqx pajBDipm asoqqi ajBunxaidd-e pni fBuuon 3IB 'sdtqsuoptpj XiioTid-BO 3ntq;B3jq puB ^qSiaAi" 'iqSiaq 'S)mpiff90o aq^ jr pado -pAap n^'^ 3<1 '^^™ ^'^^ uaAiS ■B joj pfiqo xyeiiK y -Snraren [Eois/Cqd puB noq -Dadsui [-Borpaui-jooqos q^ui naipjiqD padopAap-i]aAi ^nasaidai suuon asaqx ;g^ ;z;k^ II .S-G^ 1 .i 1 c > c C CI J C > II PHYSICAL GROWTH AND PHYSIOLOGICAL AGE 21 >^ O n H o ij w > p 1-5 -< ^ ■§ K "3) Qj a ^ W Pi o E t/2 W •suijon jnaptgaoD iqaiaq-p-jiA Du^3ni=St90o- x stajon jnapgjaoo iqSiaq-jBijA qsjiSng' sqx suiion inaptjjaoo :)qS!aq-}q2i3A\ 3iJ5aui=6iiX smiou jn'aiacyabo ;qSi3q-}q8;3Ai qsqSug aqj, sa's}!] ui' Xip^dEo Sniq^-Bsjq puE 'sureiS'o]!^ at jqSpAi 'sjaiampuaD ni ;q§i3q asTi xiiajsXs Duiaui aqi aj •;n3pg;aoD ;q3pq-:>qSp.u. S6' = m ft-^-sqi it 'sj^eaX tS joj ajdui^xa loj •juap^aoD ^q3pq-^BlIA=^qgpq-^X^I^BdB^ SaiqjBajg •^uaiDgjaoD iqSpq-5qSpM=r(qSpq4-'iq3p^\Y SI ■Bjnmjoj aqx •paj'Boipui asoqgi a}BuiixojddB pui pcajon giB 'sdrqsnopB]ai X^io'ed'Ea Suiq^^ajq pnB iqSpAi 'jqgpq 'sjiisp'tfioj aqj ji pado -pAap ipM aq XBm a3B naAiS b jo} p|iqD [[Bvas y Snrareji poisXqd puB noij -Dadsm I'Eoipaui-iooqDS q^iAi uajpuqa padojaAap-jiaAi ^uasaidaj suiuon asaqx Ed c4 2; Q 1 60 a 1 > c C 2 2 THE FIFTEENTH YEARBOOK development, and therefore that the healthy growing child should have plenty of physical and mental exercise. THE MEASURING SCALE In order that teachers, principals, superintendents, physical directors, and parents may have a basis for comparison in the study of physical development, the writer has formulated norms for physical growth in height, weight, and lung capacity, and expressed them on the accompany- ing cards (pp. 20-21), which may be used sls a. measuring 5Cfl^e for physical growth and stages of physiological maturity, as well as for records con- cerning physical conditions. The norms are high, representing the best developed children available, who have had physical training, school- medical inspection, directed play, and remedial treatment where neces- sary. A child who falls short of these standards is not necessarily subnormal physically, providing its weight and breathing capacity are proportionate to its height. The weight-height coefficients and the vital-height coefficients indicated on the card under their respective ages mark the norm to which all well-developed children should approxi- mate.^ All these norms are based upon measurements of nude children. ' Measuring scales for the growth of boys and girls have been prepared in the metric system and may be had from the writer on appUcation. An illustrated leaflet describing methods of measuring and tests for nutrition, vision, hearing, neck-glands, and posture may also be obtained from the writer. CHAPTER II NOTES ON THE DERIVATION OF SCALES IN SCHOOL SUBJECTS: WITH SPECIAL APPLICATION TO ARITID^IETIC B. R. BUCKINGHAM Chief Statistician, Department of Education, New York City The movement to set up objective means of measuring school prod- ucts, which has been so characteristic of recent educational practice, has plainly demonstrated the necessity for more and better instruments of measurement. It is in response to this need that standards of achieve- ment have been set up, and scales have been derived. The latter have been of two kinds — those based upon the judgment of competent per- sons, and those based upon the ratio of correct to total responses in a typical group. The validity of each of these bases has been questioned, but not successfully. The superiority of the one over the other has been argued, but not conclusively. In fact, there is no opposition between them. The determination of the ratio of correct responses to total responses is also, to some extent, dependent upon individual judgment, because it is frequently a matter of opinion whether a response is correct or not. In the matter of spelling, for instance, it is not always easy to say whether a given form is or is not to be scored as correct. Again, in arithmetic, which, like spelUng, is capable of relatively precise rating, it is frequently difficult to decide whether the answer to a problem should be rated as correct. In such cases readers of papers will differ somewhat in their conclusions. In other words, the element of individual judgment plays a part. Furthermore, such subjects as geography, history, and English offer much greater degrees of difficulty in the application of that preci- sion of rating upon which the ratio method of deriving scales is based. Finally, in the gradation of school subjects from the most definite subjects (spelling and arithmetic) to the least definite ones (penmanship, drawing, and English composition), we reach subjects in which the ele- ment of certainty is at a minimmn, and judgment plays its most important role. 23 24 THE FIFTEENTH YEARBOOK The point which I am seeking to make is this: that there is no oppo- sition between scales based upon judgment and those based on the ratio of correct to total responses, and that each method of derivation is most appropriate at the extremes of a series of school subjects ranging from most definite to least definite. On the score of penmanship, drawing, and EngUsh composition, a given specimen of work is superior to another, not because it is more correct, but just because people think it is superior. On the other hand, in the case of spelling, a list of words written by one child is better in point of spelling than a list written by another child, not because a mmi- ber of judges think it is the better, but because there are actually more words spelled correctly in it. In those subjects that range between the extremes of definiteness (such subjects, namely, as geography, history, and grammar) it will be evident that there is place for both the judgment and ratio methods of scale-derivation. In each of these subjects there are certain parts which are so definite and about which people differ so Uttle that there is small room for variation in judgment. In geography, for example, locational features are of this type. It is no more possible for people to difi"er about the state of the United States in which Kalamazoo is than it is for them to differ about the correct spelling of "separate" or the product of seven and nine. Similarly, in history, the question "In what 3'ear was the Stamp Act passed?" admits of no debate. And, again, in grammar, the direction "Give an example in a sentence of an adverb of time, and underline it" admits of responses which ma}?- be classified, with practical certainty, as either right or wrong. It will be noted that the questions just mentioned call for responses based on information, and it is a fact that most questions which may be rated with precision as either right or wrong are of this kind. It is obvious, however, that a large part of geography, and perhaps a still larger part of history and grammar, cannot be covered by questions of this character. Indeed, the most valuable training to be derived from these subjects is based upon the opportunity which they afford for com- parison, for inference, and for judgment. But there is no good reason why scales based upon the ratio idea should not be derived for those phases or parts of these subjects that are susceptible of the precise treatment which the ratio idea requires. In fact, it is becoming increasingly e\'ident that the present general scales THE DERIVATION OF SCALES IN SCHOOL SUBJECTS 25 will have to give place to more specific ones, each scale being suitable for a particular purpose and applicable to a given condition or situation. We shall have scales for each grade in spelling, scales for the several styles of handwriting, and scales for each type of discourse in composi- tion. Why, then, should we not have scales for information in geog- raphy, history, and grammar? The suggestion may be made at this point that in using scales based upon information, we may not only directly measure ability to give information, but we may also indirectly measure ability of a more general sort, including the power to think. Work is now being done on the prob- lem of the correlation between information and general ability within a school subject. From present indications there is reason to suppose that the correlation is high — probably not less than 0.80. If this is substantiated by sufficient data, it then becomes reasonably easy to measure something hke general ability in a given field in terms of a scale derived from questions of information. The practicability of indirect measurement is demonstrated in many of our ordinary measurements. We do not measure, for instance, the heat of the atmosphere, but, rather, the length of a mercury column, which varies directly with the amount of heat to which it is subjected. We do not measure the health of a child, but, rather, his vital index, because we either know, or think we know, that his health varies approximately with this index. When we use the phrase "varies directly," we are stating a condition which involves perfect correlation. No such perfect correlation, of course, exists between mental processes. A measure, therefore, of one process, trait, or ability in terms of another, is liable to a margin of error to the extent that the correlation is imperfect. If we may assume that the correlation between information and thought-power in geog- raphy is o . 80, we cannot get a perfect measure of " thought-power" from ''amount of information," but we can get a measure which differs from the true one by an amount small enough to be negligible for many practical purposes. With these ideas in mind, an investigation was begun in the fall of 1 9 13, which had for its purpose the development of standardized material and the derivation of scales in the subjects of arithmetic, geography, history, and grammar, upon the basis of the ratio of correct to total responses to given questions. It is to be understood that in the rating 26 THE FIFTEENTH YEARBOOK of pupils' papers, all answers were scored as either correct or not correct, and that no part credits were given. The report herewith submitted is preliminary and fragmentary. It is preliminary because it is not believed that a sufficient number of cases have been scored to make the conclusions as reliable as they should be. It is fragmentary because results are shown only for the subject of arithmetic, and within this subject, for only a portion of the questions that were used. It is offered here as an example of the extension to a wider field of a previously used method. The problems used in this report were given to seventh- and eighth- grade children in eight schools in New York City, and to children of the same grades in ten other cities. Not all the problems were used in every city. The number of participants, therefore, varies with different questions. In no case, however, was a problem attempted by less than 5,000 children. The tests were administered by the class teachers, under instructions. The principal point in the instructions was that the problems were to be written on the blackboard, one at a time, and that ten minutes were to be allowed the pupils for solving each problem. In rating the papers, children were scored as correct only in case the answer to the problem was correctly given. In a few instances, where the teacher had evidently presented an example incorrectly, and the change did not affect its nature or apparent difficulty, it was scored as correct if the pupils gave the right answer under the changed condition. In the few instances where changes were introduced by the teacher which affected the nature and difficulty of the problem, no record was taken of the results. The problems listed under Table I were given, in a preliminary test, to children in several cities other than New York City, and in a test of eight schools of a certain type in New York City, on March 23, 1915. In order to distinguish them from the problems shown under Table II, they are called the "March test." The problems listed under Table II were given, in a preliminary test, to children in a number of cities other than New York City, and were then given, also, to the same eight schools of New York City, on June 24, 19 1 5. For convenience in designation, these problems are called the "June test." THE DERIVATION OF SCALES IN SCHOOL SUBJECTS 27 TABLE I Distribution of Correct Answers by Questions and by Grades Arithmetic. March Test Seventh Grade Eighth Grade Problem First Half Second Half First H.^lf Second Half Numbers No. of Partici- pants No. Correct No. of Partici- pants No. Correct No. of Partici- pants No. Correct No. of Partici- pants No. Correct No. of Partici- pants No. Correct I 2 3 4 5 6 7 8 9 10 ....... . 1,526 1,526 1,525 1,527 1,527 1,594 1,594 1,594 1,594 1,594 143 362 681 533 544 425 816 539 751 635 1,798 1,788 1,635 1,636 1,636 1,577 1,577 1,604 1,604 1,604 201 557 898 700 801 558 985 695 833 804 1,296 1,296 1,177 1,177 1,177 1,198 1,198 1,214 1,214 1,214 308 586 785 628 669 528 869 628 753 671 1,304 1,304 1,261 1,260 1,260 1,248 1,248 1,248 1,248 1,248 520 804 963 913 885 678 985 862 821 909 5,924 5,924 5,598 5,600 5,600 5,617 5,617 5,660 5,660 S,66o 1,172 2,309 3,327 2,774 2,899 2,189 3,655 2,724 3,158 3,019 The ten problems used in the March test are the following, their numbers cor- responding to the numbers in the table above : 1. If a map 10 in. wide and 16 in. long is made on a scale of 50 mi. to the inch, what is the area in square miles that the map represents ? 2. Cream is sold in -i-pint bottles. If a milkman buys it at $1 . 20 a gallon and it costs 40 cents a gallon to bottle and deUver it, at what price per bottle must it be sold to gain 20 per cent ? 3. A fruit dealer bought 300 apples at the rate of 5 for a cent, and 300 at 4 for a cent. He sold them all at the rate of 8 for 5 cents. What did he gain on the invest- ment? 4. A family used i| bu. of potatoes a month. How much will be saved each month by buying them at $1 . 30 a bushel instead of at 8 cents a quart ? (8 qts. = I pk.; 4 pks. = i bu.) 5. James buys papers at 10 for 6 cents and sells them at i cent each. If his sales average 100 a day for 6 days, what does he add to the family income after keeping $0.10 for himself? 6. The 7A class has 42 on the roll and only 3 absent today; and the 7B class has 48 on the roU with 4 absent. Which has the better percentage of attendance and how much? 7. Bought pencils at $1 . 20 per gross, and sold them at i cent each. Find the gain per cent. 8. A family pays $25 a month for a non-heated flat and uses 5 tons of coal at $7 per ton for heating purposes during the winter. If it moved to a steam-heated flat at $30 a month, would it increase its expenses for the 3^ear or not, and how much ? 28 THE FIFTEENTH YEARBOOK 9. The value of the men's factory products in the leading centers of the United States was as follows: i860 1900 New York $17,011,370 $103,220,201 Baltimore 3,124,342 17,290,825 Boston 4,567,749 8,601,431 Chicago 540,709 36,094,310 Philadelphia 9,962,800 18,802,637 Find the increase or decrease in the value of the men's factory products in 1900 as compared with that of i860. I J. Two boys made a gallon of lemonade, using 16 lemons at 30 cents a doz. and 2 lbs. of sugar at 6 cents a poimd. They sold it at 5 cents a glass, 6 glasses to a quart. How much was each boy's share of the gain ? TABLE n Distribution of Correct Answers by Questions and by Grades Arithmetic. June Test Seventh Grade Eighth Grade Total Problem First Half Second Half First Half Second Half Numbers No. of Partici- pants No. Correct No. of Partici- pants No. Correct No. of Partici- pants No. Correct No. of Partici- pants No. Correct No. of Partici- pants No. Correct I 2 3 4 5 6 7 8 9 10 1,505 1,442 1,442 1,508 1,442 1,508 1,526 1,442 1,594 1,442 313 152 796 988 978 452 778 556 481 519 1,478 1,454 1,454 1,544 1,499 1,544 1,754 1,454 1,604 1,454 358 189 933 1,164 1,192 653 1,108 783 6x6 724 1,163 1,091 1,091 1,129 1,091 1,129 1,257 1,091 1,214 1,091 354 167 750 881 931 515 898 593 572 514 1,104 1,166 1,166 1,192 1,134 1,192 1,261 1,124 1,248 1,124 515 277 922 991 1,046 696 988 801 785 774 5,250 5,153 5,153 5,373 5,166 5,373 5,798 5,1" 5,660 5,"i I,S40 785 3,401 4,024 4,147 2,316 3,772 2,733 2,454 2,531 The ten problems used in the June test are the following, their numbers cor- responding to the numbers in the table above: 1. A farmer has a herd of 12 dairy cows that average 22 potmds each of milk per day. The milk contains 3 . 8 per cent butter-fat, and butter-fat is worth 28 cents per pound. What is the daily income from the herd ? 2. I am making a handkerchief out of a piece of Unen io| in. square. If I make a j-in. hem all around it, how long and wide will it be when finished ? 3. If Texas is 213.06 times as large as Rhode Island, and New York is 39.44 times as large as Rhode Island, then Texas is how many times as large as New York ? Express to the nearest second decimal place. THE DERIVATION OF SCALES IN SCHOOL SUBJECTS 29 4. According to the report of the Bureau of Census the numbers of persons engaged in the diEFerent groups of occupations in the United States in 1880 and 1910 were as follows: Group 1880 1910 1 7,713,875 12,567,925 2 1,871,503 7,605,730 3 3,784,726 10,807,521 4 603,202 1,825,127 5 3,418,793 5,361,033 Find the total increase in the niimber of persons employed in the United States in 1910 over those employed in 1880. 5. A can of milk containing 40 quarts costs $1 .60. What percentage is gained by seUing the milk for 6 cents a quart ? 6. I bought a cask of molasses containing 84 gallons for $28. Nine gallons hav- ing leaked out, at what price per gaUon must I sell the remainder to gain $4. 25 ? 7. A farmer's wife bought 2 . 75 yards of table linen at So. 87 a yard and 16 yards of flannel at $0.55 a yard. She paid in butter at So. 27 a poimd. How many pounds of butter was she obliged to give ? 8. A man and boy together spaded ^^^y of a garden. If the man spaded twice as much as the boy, what part of the garden did each spade ? 9. A contractor completed f of a job in 12I days. How much longer should it take to finish the job ? 10. In a certain state the cost of building a macadam road is shared by the town, county, and state. The state pays |, the coimty |, and the town the remainder. If the state pays $1,200, what does the town pay ? On the basis of the preliminary testing, the first problem in the March test was supposed to be of equal difficulty with the first problem in the June test; the second problems in both tests were, likewise, supposed to be equal in difficulty; and similarly for the remaining problems. This equality has no particular interest in the present report, and the equality is obscured by the fact that the June test was given at the very end of the school year. Tables III and IV are based upon Tables I and II, respectively, and show, for each problem and for each grade, the percentages of pupils who obtained correct answers. They also show, as a better expression of the difficulty of the problems, and, consequently, of the ability required for their solution, the equivalents of the percentages, in terms of that unit of variabiUty of the curve of error known as the "Probable Error." It is, of course, a well-known fact that there is an illusion in percentage ratings, due to the fact that the distribution of ability does not take rectangular form, but, rather, that of a logarithmic cur\^e of approxi- mately "normal" type. Account of this distribution is taken when 30 THE FIFTEENTH YEARBOOK TABLE III Percentages of Correct Answers for Each Problem and for Each Grade, with P.E. Equivalents Arithmetic. March Test Seventh Grade Eighth Grade 4 First Half Second Half First Half Second Half Per- centage Correct P.E. from Median Per- centage Correct P.E. from Median Per- centage Correct P.E. from Median Per- centage Correct P.E. from Median Per- centage Correct P.E. from Median I. . 2. . 3-- 4-- S-- 6.. 7-- 8.. 9.. lO. . • 9 • 23 ■ 44 • 34 • 35 . 26 • 51 • 33 ■ 47 • 39 4 7 7 9 6 7 2 8 I 8 +I-9S +1.06 +0.20 +0.58 +0.55 +0.92 —0.04 +0.62 +0.11 +0.38 II. 2 31.0 54-9 42.8 49.0 35-4 62. s 43-3 51-9 501 + 1.80 +0.74 — 0.18 + 0.27 +0.04 +0.56 -0.47 +0.2S — 0.07 — 0.00 23.8 45-2 66.7 53-4 56.8 44.1 72. s Si-7 62.0 SS-3 + 1.06 +0.18 — 0.64 -0.13 -0.25 +0.22 — 0.89 — 0.06 -0.4s — 0.20 39-9 61.7 76.4 72.5 70.2 S4-3 78.9 69.1 65.8 72.8 +0.38 -0.44 -1.07 -0.89 -0.79 —0.16 -1. 19 -0.74 — 0.60 —0.90 19 39 59 49 51 39 65 48 55 53 8 0 4 5 8 0 I I S 3 + 1.26 +0.41 -0.35 +0.02 — 0.07 +0.41 -0.58 +0.07 —0.21 — O.X2 TABLE IV Percentages of Correct Answers for Each Problem and for Each Grade, with P.E. Equivalents Arithmetic. June Test Seventh Grade Eighth Grade Total p« First Half Second Half First Half Second Half Ph Per- P.E. from Per- P.E. from Per- P.E. from Per- P.E. from Per- P.E. from centage Correct Median centage Correct Median centage Correct Median centage Correct Median Correct Median I. . . 20.8 + I.2I 24.2 + 1.04 30-4 +0.76 46.6 +0.13 293 +0.8I 2. . . 10. s + 1.86 13.0 + 1.67 16.6 + 1-44 23.8 + 1.06 15-2 + 1.52 3-- ■ 55-2- —0.19 64.2 -O.S4 68.7 —0.72 79.1 — 1.20 66.0 — 0.61 4-- • 65.5 -0-59 75-4 — 1.02 78.0 -115 83.1 -1.42 74-9 — 1. 00 5-. . 67.8 —0.69 79-5 — 1 .22 85-3 -1.56 92.2 — 2.10 80.3 — 1.26 6.. . 300 +0.78 42.3 +0.29 45-6 +0.16 58.4 -0.32 43-1 +0.26 7.- . 51-0 —0.04 63.2 -0.50 71-4 -0.84 78.4 -1. 17 65.6 — 0.60 8.. . 38.6 +0.43 53-9 -0.15 54-4 —0.16 71-3 -0.83 53-5 -0.13 9.. • 30-2 +0.77 38.4 +0.44 47-1 +O.II 62.9 -0.49 43-4 +0.25 10. . . 36.0 +0-53 49.8 +0.01 41. 1 +0.33 68.9 -0-73 49.1 +0.03 THE DERIVATION OF SCALES IN SCHOOL SUBJECTS 31 percentage values are thus converted into some unit of variability of the curve of error. ^ The twenty problems listed in Tables I-IV, as has been stated, were given in March and in June, 19 15, to typical schools in New York City. On the basis of the returns received from these schools, the figures given in Tables V and VI were made up. These figures refer to the scores of individuals. As measures of the ability of groups of children, these tables may prove useful. They are, however, only approximations. TABLE V Distribution of Pupils according to the Nximber of Problems Answered Correctly Arithmetic. March Test No. OF Problems Correct Grade Vri' Grade VII' Grade VIII' Grade VIII" Grades VII" to VIII» No. Percent- age No. Percent- age No. Percent- age No. Percent- age No. Percent- age 0 I 2 3 4 5 6 7 8 9 10 190 230 160 204 161 156 114 98 77 43 9 13.2 16.0 II. I 14. 1 11. 2 10.8 7-9 6.8 S-3 30 0.6 144 157 147 1 54 178 147 150 126 121 107 9 8.1 II. I 10.4 10.9 12.6 10.4 10.6 8.9 8.6 7.6 0.6 43 73 88 96 112 132 131 136 125 96 59 3-9 6.7 8.1 8.8 10.3 12. 1 12.0 12.5 II-5 8.8 5-4 6 18 42 51 67 115 117 160 163 189 114 0.6 1-7 4.0 4-9 6.4 11. 0 II-3 154 15-6 18. 1 10.9 353 478 437 505 518 550 512 520 486 435 191 7-1 9.6 8.8 10. 1 10.4 II. 0 10.3 10.4 9.8 8.7 3-8 Total Median . . . 1,442 2.691 1,410 4-747 1,091 6.015 1,042 7.656 4,985 5-367 The fact that the ten problems upon which each of these tables is based varied widely in point of difficulty makes conclusions as to individual abilities somewhat unreliable. Tables made up from material each part of which was of equal difficulty would be more reliable, but no such tables have ever been constructed, for the reason that no material meeting these conditions exists. Meanwhile, therefore, we shall be obliged to content ourselves with the usual method of approximation. With this caution * For conversion tables see E. L. Thomdike, Mental and Social Measurements, 2d ed., p. 228; also B. R. Buckingham, Spelling Ability, Its Measurement and Distribution, p. 116. 32 THE FIFTEENTH YEARBOOK in mind, Tables V and VI may be used as standards of attainment. Figs. 1-8 show, in graphic form, the frequencies of each rating as given in these tables. TABLE VI Distribution or Pupils according to the Number of Problems Answered Correctly Arithmetic. June Test No. OF Grade VII' Grade VII" Grade VIII' Grade VIII» Grades VII' to vni" Correct No. Per- centage No. Per- centage No. Per- centage No. Per- centage No. Per- centage o I 2 3 4 5 6 7 8 9 lO 67 144 193 23s 211 204 143 97 70 54 24 4.6 10. 0 134 16.3 14.6 14. 1 9.9 6.7 4-9 3-7 1-7 28 6S 95 165 200 243 204 189 134 69 18 2.0 4.6 6.7 II. 7 14.2 17.2 14-5 134 9-5 4-9 1-3 13 32 71 III 171 167 146 142 129 79 30 1.2 2.9 6.5 10.2 15-7 15-3 13-4 13.0 II. 8 7.2 2.7 4 4 27 62 73 116 145 184 159 189 79 0.4 0.4 2.6 6.0 7.0 II. I 139 17.7 153 18. 1 7.6 112 245 386 573 655 730 638 612 492 391 151 2.2 4-9 7-7 II-5 I3-I 14.6 12.8 12.3 9-9 7-8 3-0 Total Median . . . 1,442 4-389 1,410 5.626 1,091 5.886 1,042 7.489 4,985 5-715 Fig. 9 shows, for each grade, in the form of a scale, the facts with respect to the difiSculty of the problems that were shown in the columns headed " P.E. from Median " in Tables III and IV, although the problems of the June test are not strictly comparable to those of the March test, on the basis of the returns received, for the reason that there is an increment of ability among school children during a three months' period. The two sets of problems are scaled on the same projection, but they are kept separate by showing, in Fig. 9, the numbers of the March problems above the scale line, and the numbers of the June problems below the scale line. Fig. 9 also shows a general scale for the four grades combined; the point of reference is the median of Grade VII^ In order thus to refer the results of the testing in higher grades to the median of the lowest grade, it is necessary to know the intervals between the medians of the successive grades. THE DERIVATION OF SCALES IN SCHOOL SUBJECTS 33 Tables VII and VIII show, for the March and June tests, the number of pupils in each grade who equaled or exceeded the score of the median TABLE VII Amount and Percentage of Overlapping with P.E, Equivalenxs Arithmetic. March Test Grade VH- Grade VII» Grade Vm- Grade VUI' Grade VIP No. Per cent P.E. No. Per cent P.E. No. Per cent P.E. No. Per cent P.E. 538 37-3 +0.48 340 25-3 +0.99 5" 36.2 +0.52 163 "•3 + 1 80 Grade VIP 1038 73-6 -0.94 915 83 -9 -1.47 989 950 -2.44 281 19.9 + 1.25 327 30.0 +0.78 Grade VIII' 708 64.9 -0-S7 875 84.0 -1.48 Grade VHP 742 71.2 -0.83 Table reads: 538 pupils of Grade VII" equaled or exceeded the score of the median pupil of grade VII', which was 37.3 per cent of all pupils of Grade VII'. The equivalent of this percentage is 0.48 P.E., etc. TABLE VIII Amount and Percentage of Overlapping with P.E. Equivalents Arithmetic. June Test Grade VII' Grade VII' Grade VHI' Grade VIH' Grade VIP No. Per cent P.E. No. Per cent P.E. No. Per cent P.E. No. Per cent P.E. 465 32.2 +0.69 412 28.6 +0.84 642 45-5 +0.17 198 13-7 + 1 62 Grade VIP 980 69-5 — 0.76 789 73 I —0.91 917 88.1 -1-75 318 22 6 +1.12 Grade Vnii 589- 540 -O.IS 800 76.8 — 1.09 3" 28.5 +0.84 Grade VIII' 770 73-9 -0-95 Table reads: 465 pupils of Grade VII' equaled or exceeded the score of the median pupil of Grade VII', which was 32.2 per cent of aU pupils of Grade VII'. The equivalent of this percentag is 0.69 P.E., etc. 34 THE FIFTEENTH YEARBOOK 567 Fig. I 9 10 ^ 01 23 456789 10 Fig. 2 Figs, i, 2, 3, and 4. — Frequency of each rating (of problems correct) in Gra for problems answered correctly. The vertical scale is for the percentage of child i Grade 7^; 1,091 for Grade 8'; and 1,042 for Grade 8*. THE DERIVATION OF SCALES IN SCHOOL SUBJECTS 35 IS 123456789 10 Fig. 4 ■, 7', 8', and 8^ respectively. March test. (See Table V.) The horizontal scale is ho obtained each correct number of answers. N = 1,442 for Grade 7'; 1,410 for 36 THE FIFTEENTH YEARBOOK IS E ^ -\ . 1 ' 1 S 6 Fig. s 9 lo IS 5 6 Fig. 6 lO Figs. 5, 6, 7, and 8. — Frequency of each rating (of problems correct) in Grad for problems answered correctly. The vertical scale is for the percentage of childn Grade 7'; 1,091 for Grade 8'; and 1,042 for Grade 8*. THE DERIVATION OF SCALES IN SCHOOL SUBJECTS 37 c£: X u 0123456789 10 Fig 7 012 3456789 10 Fig. 8 7', 7', 8', and 8*, respectively. June test. (See Table VI.) The horizontal scale is who obtained each correct number of answers. N= 1,442 for Grade 7'; 1,410 for 38 THE FIFTEENTH YEARBOOK pupil of each of the other grades. They also show the percentages and P.E. values corresponding to these numbers. The P.E. values are the intervals between the grade medians. These tables are constructed upon the assumption of a normal distribution of ability in all grades, and upon the further assumption that the variability in any one grade is equal to the variability in each of the other grades. Obviously, we have several expressions for the same relationship; for instance, in Table VIII, the distance between the median of Grade VII^ and the median of Grade VII^ is (column 4) 0.69. In the same table, the same distance measured in the opposite direction is 0.76. Two similar values for the same distance are shown in Table VII; and besides these four measures, a number of others may be derived. Using, how- ever, only direct statements of the relationship between consecutive medians, we have, in each case, four quantities. The averages of these, being taken, give the following results: Median of Grade VIP to median of Grade VIP =0.72 P.E. " " " VIP " " " " VIIP = o.37P.E. " " VIIP " " " " VIIP = o.8sP.E. To obtain the general scale shown in Fig. 9, all that is necessary to do now is to add to the P.E. values of Tables III and IV, for grades higher than VII', the interval at which the medians of these grades stand above the median of Grade VII'. By so doing, three values in addition to the one for Grade VII' will be found, and the average of these four may be taken as the best position at which to "place" the problem in question. Table IX gives the distance at which each problem stands above the median of Grade VII' when computed on the basis just described. The general scale in Fig. 9 is the graphic representation of this table. It is clear that the scales derived in this paper are very meager, and, as was said in the beginning, this report is merely preliminary and suggestive. A far greater number of problems in arithmetic should be used for the purpose of constructing a more complete scale. In fact, 120 such problems are now being worked up with this end in view. Material is likewise in hand for a large number of questions in geog- raphy, history, and grammar, which will be scaled in the same way. THE DERIVATION OF SCALES IN SCHOOL SUBJECTS 39 i »- 6— ■4 «» 9 » a- « • < *-- ♦ 1) ■a $ o d O >>- V •> a- 9 1 u e 1 *i s, o °3 S S o (^ -3 40 THE FIFTEENTH YEARBOOK Any superintendents or principals who are willing to give tests in these subjects are invited to communicate with the writer of this chapter. TABLE IX Average Position of Problems, Ex- pressed AS Distances from the Median of Grade VII and in Units OF P.E. Problem March June I 2.24 1.32 0.52 0.90 0.83 1.32 0.29 0.96 0.69 0.76 1.72 2.45 0.28 2 7, 4 — O.II c —0.46 6 1. 17 7 0.30 0.76 8 0 I -IS 0.97 lO CHAPTER III SCORE CARD FOR CITY SCHOOL BUILDINGS GEORGE DRAYTON STRAYER Professor of Educational Administration, Teachers College, Columbia University The score card which is printed herewith has been developed as a part of the advanced work in educational administration under the direction of the author of this article, with five different groups of grad- uate students, through two academic years and one summer session.^ The idea of a score card has been common over a considerable period of years, especially in the work of agricultural colleges. There is a mani- fest advantage in the score card in that it fixes attention upon all of those qualities or elements which go to make up the perfect whole desired. Individuals in judging school buildings not infrequently think mainly in terms of two or three elements which seem to them to be of primary importance, and often neglect other parts of the building which are, when one stops to consider them, of equal value. In making the score card, it has been necessary first of all to include as nearly as possible all those details which go to make up the perfect school building. It was, of course, desirable, in so far as it was possible, to include under a few main heads all the subordinate factors. It was only after a very considerable amount of experimentation that the heads "Site," "Build- ing," "Service Systems," "Classrooms," and "Special Rooms" were decided upon. After organizing the score card in terms of the large and the sub- ordinate heads, the next step was, of course, to assign to each main and each subordinate head the proper weight out of a total of a thousand points which was allowed as a perfect score. The method employed was ' Special acknowledgment is due to Messrs. L. H. King, B. W. Loomis, and A. Dushkin, who were constituted as the first committee to draft a score card for school buildings, and whose work was used as a basis for discussion in the subsequent development of the form which is here printed. The author is indebted to Dr. M. R. Trabue for the statistical calculations necessary to determine the weights to be assigned to each of the several items. 41 42 TEE FIFTEENTH YEARBOOK to ask experienced superintendents and principals of schools taking work in educational administration, which involved a considerable knowledge of statistics, to assign to each element the weight which they thought should be attached. The five large heads were scored first, next the several main subdivisions imder each large head, and finally the elements under the last-named divisions. More than two hundred students participated in this part of the work. From the ratings assigned by the last group of a hundred students working upon the score card, after studying the results which had already been secured by the former groups, medians were calculated, first, for the large heads, and then for each of the subordinate heads. The medians were calculated only to the nearest five points on the scale. This means that the score card as it appears represents the combined judgment of the scorers to within two and one-half points. In many cases, of course, the median fell on five or some multiple of five, in which case the score given corresponds exactly to the median found. The method employed in determining the weight to be assigned to each element appearing on the score card needs no particular defense, since the value of any particular part or element in the construction of the building is, after all, a matter of judgment. The median judgment derived from the scores allowed by a large group of those competent to judge of the worth of the several elements is the best single measure which can be found. In order to use the score card, one should be familiar with the prob- lems involved in schoolhouse construction. To that end, there appears at the end of the score card a bibliography. It would be well, in training people to use the score card, to have them thoroughly conversant espe- cially with the more important authorities on the subject of school hygiene and school architecture. After such study, visits to buildings, under the guidance of some competent student of these problems, would add greatly in training persons to use the card. The card is given first in a very brief and highly condensed form, which could be used only by those entirely familiar with the longer form and with the meaning of each of the headings which is there listed. As a matter of practice, it will be best first to use the longer form of the score card, and, only, after considerable facility has been acquired, to drop the longer form in favor of the shorter summary. For one familiar with school buildings and with the score card, much would be gained by checking over blueprints SCORE CARD FOR CITY SCHOOL BUILDINGS 43 and specifications in the light of the score card before beginning to con- struct the building. Here again the value is in large measure to be found in the fact that each of the more important items will be brought to the attention of the one who seeks to criticize the plans and specifi- cations, and their relative importance will at least in some measure be indicated. It will be found particularly worth while to score old buildings, in order to call attention to the necessity for reconstruction which is always to be found in a city in which buildings have been in use over a consider- able number of years. As one studies the problem of school buildings in the United States, he is impressed by the accidental or occasional repair or reconstruction which is provided. A careful study and scoring of these buildings will often indicate common deficiencies of very great importance which should receive immediate attention, and others which are of relatively less significance which may be postponed for a time. In the same school system it may be found as well that one building is so remarkably more deficient than another that it is manifestly good pubUc policy to spend whatever money is available in reconstructing the buildings which scores lowest before undertaking the work which may not be nearly as important in other buildings. In the case of scoring school buildings, as with any other instrument of measurement, the result should suggest problems, and in some measure indicate the direction in which reforms are to be brought about. Any person using the score card should supplement the mere scoring of the several items with a report upon any notable deficiency which renders the buUding unfit for use. It is entirely conceivable that a building on most counts might stand high, but in some one particular, say with respect to fire protection or sanitation, might rate as entirely unsatis- factory. In this case the notation, after the building was scored, would call attention to the fact that measures should be taken immediately to remedy particular defects, in which case the building would, possibly with a minimum of expense, be brought up to a very high standard of excellency. The writer will be very glad to receive, from anyone who may use the score card, criticisms or suggestions for its improvement. It will be particularly helpful to receive reports of the use of the score card, showing to what degree two or more individuals scoring the same build- ing arrive at the same result. 44 THE FIFTEENTH YEARBOOK SCORE CARD FOR SCHOOL BUILDINGS AND EQUIPMENT FOR CITY ELEMENTARY AND HIGH SCHOOLS City Building Principal Date Enrolment: Boys Girls Total Average Daily Attendance: Boys Girls Total Number of Rooms Approximate Cost Scorer Instructions 1. Abbreviation: S — standard. 2. Basis for scoring — i,ooo points. 3. In scoring classrooms, stairways, entrances, fire escapes, and the like, score each separately and insert the average for the final score. 4. It will be worth while to use this card in checking up blueprints of prospective bmldings. To do this will require a complete set of specifications with the blue- prints, also a copy of state laws and city ordinances. SHORT FORM OF SCORE CARD I. Site (125) A. Location (55) I. Accessibihty (25) 2. Environment (30) B. Drainage (30) I. Elevation (20) 2. Nature of Soil (10) C. Size and Form , (40) n. BtnxDiNG (165) A. Location (25) I. Orientation (15) 2. Position on Site (10) B. External Structure (60) I. Type (5) 2. Material (10) 3. Height (5) 4. Roof (5) 5. Entrances (10) 6. Aesthetic Balance (10) 7. Condition (15) C. Internal Structure (80) I. Stairways (35) 2. Corridors (25) 3. Basement (15) 4- Attic (5) III. Service Systems (280) A. Heating and Ventilation System (70) I. Kind (20) 2. Installation (10) 3. Air Supply (25) 4. Distribution (15) B. Fire Protection System (65) I. Apparatus (10) 2. Fireproof ness (20) 3. Escapes (20) 4. Electric Wiring (s) 5. Fire Doors (10) J SCORE CARD FOR CITY SCHOOL BUILDINGS 45 C. Cleaning System (20) D. Artificial Lighting System (20) E. Electric Service Systems (15) I. Clock (s) 2. Bell (s) 3. Telephone (5) F. Water Supply System (30) G. Toilet System (50) I. Distribution (10) 2. Fixtures (10) 3. Adequacy (10) 4. Seclusion (5) 5. Sanitation (15) H. Mechanical Service Systems (10) I. Elevators (5) 2. Book-Lifts (2) 3. Waste-Chutes (3) IV. Classrooms (290) A. Location and Connections (35) B. Construction and Finish (90) I. Size (25) 2. Shape (15) 3, Floors (10) 4. Walls (10) 5. Doors (5) 6. Closets (5) 7. Blackboards (10) 8. Color-Scheme (10) C. Illumination (85) I. Glass Area (45) 2. Windows (30) 3. Shades (10) D. Cloakrooms and Wardrobes (25) E. Equipment (55) I. Seats and Desks (40) 2. Teacher's Desk (10) 3. Bulletin Boards (5) V. Special Rooms (140) A. Large Rooms for General Use (65) I. Pla5T:oom (10) 2. Auditorium (15) 3. Study-HaU (5) 4. Library (10) 5. Gymnasium (15) 6. Lvmchroom (10) B. Rooms for School Officials (35) I. Offices (10) 2. Teachers' Room (10) 3. Nurses' Room (10) 4. Janitor's Room {<,) C. Other Special-Service Rooms (40) I. Laboratories (20) 2. Lectiire-Rooms (10) 3. Storerooms (5) 4. Studios (s) DETAILED SCORE CARD FOR CITY SCHOOL BUILDINGS I. Site A. Location: 1. Accessibihty — centrality (present and future), car lines, streets. 2. Environment: o) Physical — gardens, trees, shrubbery, buildings, hills. (S — Skyline should not have an angle of more than 30 degrees from base of building.) b) Social — density of settlement, composition, moral influences. c) Protection — freedom from noise, dust, danger, malodors. 46 THE FIFTEENTH YEARBOOK B. Drainage: 1. Elevation, slope. (S — Grounds should slope away from building and should not exceed i in. for every 3 ft.) 2. Nature of soil — residual or artificial, kind, texture, aeration, hydration, surface material. C. Size and Form: Should be large enough and of good shape to allow for proper placing of buildings, for 30 sq. ft. of playground per child, and for school garden. II. Building A. Location: 1. Orientation — light, exposure. (S — Southeast, east, southwest, west, and south in order.) 2. Position on site as regards appearance and economy of playgrounds. B. External Structure: 1. Type — rectangle, square, inner court, T, H, E, U. 2. Material. (S — Brick or stone.) 3. Height — number of stories. (S — Two stories above basement.) 4. Roof — type and material. (S — Flat, waterproof, suitable for playground, proper slope for drainage.) 5. Entrances: a) Number, location width. (S — At least two, near stair landings, 11-13 ft, wide.) b) Steps — munber, protection from the elements. (S — As few as possible, unexposed.) c) Vestibules — size, lighting. (S — 11-13 ft. wide, double-swing glass doors, and waterproof floors.) d) Doors — kind, opening, springs, checks, stops. (S — 3I ft.X8 ft., opening outward with panic bolts.) 6. Aesthetic balance. (S — SimpHcity and utiHty.) 7. Condition. (S — Should be in good repair.) C. Internal Structure: I. Stairways. o) Construction — kind (box, open, winding), material, tread and riser, nosing, width, landing, banister (number, kind, size, stabUity), soundproof ness. (S — Tread, 11-13 in.; riser, 7 in.; width, 5 ft.; metal banister, i| in. dia., at least two for each stairway, firmly attached.) h) Number and location — proximity to exits. (S — At least two, landings near exits.) c) Lighting — natural and artificial. (S — Should be well Ughted.) d) Sanitation — coves, comers, ledges. (S — Should have sanitary coves and be free from dust-catchers.) SCORE CARD FOR CITY SCHOOL BUILDINGS 47 2. Corridors. a) Location. b) Construction — material, width, door arrangement, finish (chair rail, picture mold, dado). (S— Width 11-13 ft.) c) Obstructions — blockers, cases, pedestals. (S — These should not obstruct easy passage.) 3. Basement. a) Depth below grade, dampness, areas. (S — Depth, 3 ft.; floor and walls waterproof.) b) Boiler-room, accessibility to fuel-room, exits, ash-lifts. c) Fuel-room, size, construction, chute. 4. Attic, waterproof, heatproof, floor. in. Service Systems Note. — Defects in any service system should be checked against the system, wherever found. A. Heating and Ventilating System: 1. Kind of system — direct, direct-indirect, gravity, plenum, plenum-exhaust. 2. Installation — piping, workmanship, noise, control. (S — All piping should be insulated.) 3. Air supply — source, amount, humidification, ducts. (S — From the top of the building; humidity 40-60 per cent; 2,000 cu. ft. per hour per pupil, should not enter with a velocity greater than 6 ft. per second.) 4. Distribution — size, arrangement, kind of ducts, pipes, and radiators. (S — Single ducts for each room; inlets 8-9 ft. above floor, outlets near floor.) B. Fire Protection System: 1. Apparatus — fire hose, extingmshers, water pressure, fire alarms. (S — Adequate supply on each floor; fire alarms easily accessible, automatic in boiler-room, connected with city fire department.) 2. Fireproof ness: a) Building as a whole — rating of underwriters. b) Stairways. (S — Encased fireproof stair-wells.) c) Boiler- and fuel-rooms. (S — Separate fireproof rooms.) 3. Fire escapes — number, location, kind, protection, number of exits. (S — In non-fireproof buildings there should be at least two fire escapes.) 4. Electrical work — nature and place of intake, insulation, number and kind of outlets, location of switches, meter, cut-out, cabinets. (S — Should be installed according to rules of underwriters.) 5. Fire doors — kind, location, operation. (S — Automatically closing.) C. Cleaning System: Kind, installation, efficiency. (S — Vacuum system.) D. Artificial Lighting System: Kind, amoimt, distribution, number, and location of switches, outlets for lanterns, etc. 48 THE FIFTEENTH YEARBOOK E. Electric Service Systems: 1. Clocks. 2. Bells and gongs. 3. Telephones — number and location. (S — At least one on each floor.) F. Water Supply System: Drinking-fountains, baths, lavatories, janitor's supply (on each floor). Installation and sanitation. G. Toilet System: 1. Distribution — location, accessibility. (S — Most on first floor, others dis- tributed.) 2. Fixtures — seats, urinals, washbowls, sinks, towel and paper holders — size, kind, durability, and arrangement. 3. Adequacy — number. (S — i seat for each 15 girb, i seat for each 25 boys, I urinal stall for each 10 boys.) 4. Seclusion — soimdproofness, doors. 5. Sanitation — finish, material, workmanship, lighting, ventilating. fS — Mate- rial— not absorbent, non-corrosive.) H. Mechanical Service Systems: 1. Elevators (for buildings of more than four stories) — location, fireproofness, adequacy. 2. Book-Ufts. 3. Waste-chutes — kind, location, size. (S — Fireproof, outlets closing auto- matically.) rv. Classrooms A. Location and Connections (to exit, drinking-foimtains, toilet). Deduct for base- rooms and those above fourth floor without elevators. B. Construction and Finish: 1. Size. (S — Per pupil 15 sq. ft. floor space and 200 cu. ft. air space.) 2. Shape — classroom 24X30X12 ft. 3. Floors — material, condition (cracks, checks, splinters, loose boards, projecting ends) , width of boards, soimdproofness, cove, baseboard, surface, finish. Stone, tile, cement, and other composition floors are bad for class- or study-rooms. (S — Should be battleship-linoleum, or hard wood, durable, well joined, and not dust-retaining.) 4. Walls, ceUing — plastering, finish, texture, condition, picture mold, chair rail, kind and condition of dado. (S — Hard, smooth, non-glass plaster, with cement plaster for dado, avoiding grooves and ledges.) 5. Doors — how opened, size, kind, lock, threshold, transom, number of exits. (S — Doors without thresholds and transoms.) 6. Closets and bookcases — location, size, convenience. 7. Blackboards — kind, length, width, color, chalk rail, height from floor, surface, quaUty, condition, trim. (S — Slate, full black, on front and side of room, 36-42 in. wide, height of chalk rail, grades I-II, 24 in.; III-IV, 26 in.; V-VI, 28 in.; VII-VIII, 30 in.; high school, 32-36 in.) SCORE CARD FOR CITY SCHOOL BUILDINGS 49 8. Color-scheme — woodwork, dado, walls, ceiling, furniture, shades, finish, fix- tures. (S — Neutral color, buff or green; dado slightly darker than walls, white or cream ceiling; woodwork, furniture, and shades to harmonize in tone; dull, smooth finish.) C. Illumination: 1. Glass area — 5- to | area of floor. 2. Windows — size of mullions, nearness to ceUing, height of sill, kind of glass, distance of front window from front wall, orientation, shape. (S — From pupils' left, imilateral, grouped, symmetrical, as near ceUing as possible, 3I to 4 ft. from floor, plain glass, mullions not over 1 2 in. wide. Front windows should not come within 5 ft. of front waU; easterly exposure best; rectangxilar in shape.) 3. Shades — kind, material, hanging, adjustment, condition. (S — Adjustable from center.) D. Cloakroom, Wardrobes: Location, size, convenience, ventilation, finish. (S — Ample ventilation and accommodation.) E. Equipment: 1. Seats and desks — kind, number. (S — Adjustable-movable or adjustable; not over 42 in niunber.) 2. Teacher's desk. (S — No platform.) 3. Bulletin boards. V. Special Rooms A. Large Room for General Use: 1. Playroom — location, size, accessibility, adaptabihty, finish. (S — Per pupil IS sq. ft. floor space and 200 cu. ft. of air space.) 2. Auditoriimi: a) Location, accessibility. (S — Should be on first floor.) b) Construction — size, height, seating capacity, floor, acoustics, exits, obstruc- tions, gallery (kind, seating capacity, location), light and natiire of stage, finish, ornamentation. (S — For 80 ft. length on flat floor, stage should be 3 ft. 8 in. high; on dish floor, 3 ft.) c) Auxiliaries — dressing-rooms, curtain, setting, seats (kind, arrangement). 3. Study-hall — location, size, accessibility (especially to library), adaptabihty, finish. 4. Library — location, size, accessibiUty, form and arrangement of stacks. 5. Gymnasium: a) Location — accessibility, segregation of sexes. b) Construction — size, floor, track, gallery, soundproofness, finish. (S — Height 22-25 ft. Length and width should relate as 3 to 2.) c) Auxiharies — lockers, showers, dressing-rooms (number, kind, location, con- venience, condition.) 6. Limchroom — location, accessibiUty, size, adaptability, arrangement, finish, sanitation. 50 TEE FIFTEENTH YEARBOOK B. Rooms for School Officials: 1. Ofl&ces — location, size, adaptability, finish; waiting-room (ditto). 2. Teachers' room — location, size, toilet facilities, equipment, finish. (S — Equip- ped with chairs, couch, gas or electric plate.) 3. Nurses' room — location, size, equipment and toilet facilities (including bath) adaptabiht}', sanitation, finish. 4. Janitor's room — location, size, convenience. C. Other Special-Service Rooms: 1. Laboratories: Note. — Include all facilities for chemistry, physics, biology, physiography, commercial work, household and industrial arts. a) Kind, location, size, adaptability. (S — Depends on nimiber of pupils to be accommodated. A room 30X40 ft. will accommodate 25 pupils.) b) Construction — plimibing, storerooms, cabinets, finish. 2. Lecture-room — location, size, seating capacity, plumbing facilities, accessibility, fixed furniture (number, kind, arrangement). 3. Supply- and storerooms — location, size, adaptability. 4. Studios — kind, location, size, and adaptability. Note. — Include drawing-, art,- and music-rooms. BIBLIOGRAPHY 1. Barry, W. F. The Hygiene of the Schoolroom (1903). 2. Briggs, W. R. Modern American Houses (1899). 3. Bruce Publishing Company. School Architecture (Handy Manual) (1910). 4. BtJRGERSTEiN, Leo. School Hygiene (Trans.) (1915). 5. BtJRGERSTEiN, Leo, AND Netolitzky, A. Handbuch der Schulhygiene (1902). 6. Burrage, S., and Bailey, H. T. School Sanitation and Decoration (1899). 7. Carpenter, A. The Principles and Practice of School Hygiene (1887). 8. Carpenter, R. C. Heating and Ventilating Buildings (1896). 9. Clay, Felix. Modern School Buildings (1906). 10. Dresslar, F. B. American Schoolhouses (U.S. Bull. No. 5) (1910). 11. Dresslar, F. B. School Hygiene (1913). 12. Elkington, J. S. C. Health in the School (1907). 13. Ellis, A. C, and Kuehne, H. School Buildings (Bull. Univ. of Texas) (1905)- 14. Feret, a. L'Hygiene scolaire (1900). 15. Hamlin, Snyder, and Others. Modern Schoolhouses (19 10). 16. Hope, E. H., and Browne, E. A. A Manual of School Hygiene (1901). 17. Kidder, F. E. Architects and Builders Pocketbook (1892). 18. Kotelmann, L. School Hygiene (Trans.) (1899). 19. Mills, W. T. American School Building Standards (1910). SCORE CARD FOR CITY SCHOOL BUILDINGS 51 20. Moore, J. A. The School House (1905). 21. Putnam, H. C. School Janitors (1913). 22. Rapeer, L. W. Educational Hygiene (1915)- 23. RowE, S. H. The Lighting of Schoolrooms (1904). 24. Shaw, E. R. School Hygiene (1902). 25. Wheelwright, E. M. School Architecture (1901). 26. Whipple, G. M. Questions in School Hygiene (1909). 27. Young, A. G. School Hygiene and Schoolhouses. (Seventh Annual Rept. State Board of Health, Augusta, Maine, pp. 83-385) (1892). CHAPTER IV COMPLETION TESTS FOR PUBLIC-SCHOOL USE M. R. TRABUE Instructor in Educational Administration, Teachers College, Columbia University Many of the tests upon which psychologists depend for their knowl- edge of an individual's mental characteristics are of such a nature that it is almost impossible to make extended use of them in the schoolroom. There are other very reliable psychological tests, however, which may be adapted to schoolroom use with comparatively few changes. One of these is the completion test, which psychologists have come to regard as an unusually good test of ability to think about words and language forms. In view of the fact that so much of the child's school work is dependent upon his ability to read and interpret printed words, it has seemed worth while to the writer to make such changes in the form of the completion test as will make it available for general use in schools.^ The scale shown below will serve as an example of the proposed new completion-test forms. It is believed that these new forms will be found very helpful to school officers in measuring the abilities of children and in classifying them accordingly. The foregoing form is designed to meet the two or three most serious obstacles which heretofore have confronted the school 'officer who wished to employ the completion test in his school system. In the first place, the forms commonly used by the psychologists cannot be employed in the middle and lower grades of the elementary school because they are too difficult. Ebbinghaus, who was the first to employ the completion-test method, used mutilated paragraphs. In this respect he has been fol- lowed very closely by later investigators, and in almost every instance the incomplete paragraphs have been too difficult for practical use with elementary school pupils. I have attempted, in preparing forms for school children, to use the sentence, rather than the paragraph, ' M. R. Trabue, "Completion-Test Language Scales," Teachers College Contribu- tions to Education, No. 77, New York, 1915. 52 COMPLETION TESTS FOR PUBLIC-SCHOOL USE 53 as the unit of thought. I have found it possible to begin a test with sentences so simple that a large majority of the second-grade pupils are able to complete them correctly, and to finish the test with sentences so difficult that only a small percentage of the Freshmen in college com- plete them. Write only one word on each blank Time Limit : Seven minutes Name Trabue Language Scale B I. We like good boys girls. 6. The is barking at the cat. 8. The stars and the wUl shine tonight. 22. Time often more valuable money. 23. The poor baby as if it were sick. 3 1 . She if she wUl. 35. Brothers and sisters always to help other and should quarrel. 38 weather usually a good effect one's spirits. 48. It is very annoying to tooth-ache, often comes at the most time imaginable. 54. To friends is always the it takes. A second difficulty with previous completion-test forms arises from the fact that we do not know the relative values of the various commonly used paragraphs, which makes it practically impossible to measure progress from year to year or from grade to grade. Professor Whipple emphasizes this point, as follows: "Since the ehsion of a single letter may, in some circumstances, very considerably increase the difficulty of the test, it follows that, without extensive preliminary trials, it is well- nigh impossible to prepare a series of texts of equivalent difficulty, or to insure that the several sections within a given text present equivalent difficulty."^ This objection has been met by actually trying the incomplete sentences upon thousands of public-school children and discovering from the results just how difficult each sentence is for each class of children, and for all children together. From the results thus obtained, ' G. M. Whipple, Manual of Mental and Physical Tests, Part 11 (1915), p. 284. 54 TEE FIFTEENTH YEARBOOK four approximately equal scales have been derived (Scale B, shown above, being one of the four), each scale consisting of ten sentences, which are arranged in the order of their difficulty from simple to hard. By measuring ability at the beginning of a year with one scale and then at the beginning of the next year with an equivalent scale, it is possible by subtracting the first result from the second to determine the amount of change effected in a class or in a child during a year. A third difficulty with the paragraph form of the completion test is found in the scoring. Ebbinghaus scored according to the number of syllables correctly supplied, but this method is inadequate, for some syllables are ten times as hard to supply as others. Later investigators have estimated the quality of the completed paragraphs as wholes, giving loo per cent for a perfectly completed paragraph, 50 per cent "if the inserted words make a well-connected story, but related in only a moderate degree to the thought that should have been given," and no credit at all for words which are "purely literary invention, having no connection with the thought given by the printed words." It will readily be seen that very careful consideration and judgment are required if one is to assign accurate and comparable scores in the foregoing manner. Even trained psychologists have difficulty in agree- ing just how much a given completion is worth. Teachers and school administrators are usually too busy with other school problems to spend much time in such tiresome mental labor as is required to assign scores to partially completed paragraphs. After making an attempt to distinguish six grades of quality (5-4-3- 2-1-0) in the completion of a sentence,^ the writer found that practically nothing was lost by simplifying the scoring still further, giving 2 points credit for each perfectly completed sentence, i point for each sentence completed with only a slight imperfection, and o for any sentence omitted or imperfectly completed. It is infinitely easier and quicker to say that a sentence is "right," or that it is almost "right," than it is to decide that a partially completed paragraph is worth 78 per cent of a perfectly completed paragraph. The writer has also published in the appendk of his report the detailed scheme by which each individual sentence was scored, in order that teachers or administrators who do not wish to be I J. L. Stenquist, E. L. Thomdike, and M. R. Trabue, "The Intellectual Status of Children Who Are Public Charges," Archives of Psychology, No. 33, September, 1915, pp. 13-19. COMPLETION TESTS FOR PUBLIC-SCHOOL USE 55 bothered with the making of judgments as to whether a sentence is right or wrong may have before them very objectively just what has been called "right" and what has been called "wrong." The writer is rather firmly convinced that ability to perform mental tasks can be measured more adequately by a graded series of perform- ances to be done in a given time than by any other scheme. The measurement of physical ability to lift weights at arm's length may be taken as an analogy. If we had a series of ten weights, ranging from 20 to 200 pounds by steps of 20 pounds between consecutive weights, as represented in Fig. i, we might measure a boy's ability by having him begin at the lightest of the series and at arm's length lift in their order to the level of his shoulder as many of the weights as possible. In order to make the analogy with Scale B complete, we should have to give 2 Fig. I. — Representing a series of graduated weights. points credit for each weight lifted to a level with his shoulders, and I point for each weight which was lifted "almost but not quite" to the shoulder level within a time-limit of seven minutes. With the older completion-test forms the analogous measurement would be as follows: give the individual the 200-pound weight and see how nearly to the shoulder level he can raise it in ten minutes. It is clear at once that a test arranged on such a plan could be used for only a short fraction of the total range of ability, and that scores assigned by dififerent individuals for the same quality of performance would vary considerably. With certain commonly used tests in arithmetic the analogous measurement would be somewhat as follows: give the boy a large number of 100-pound weights and see how many of them he can at arm's length lift to shoulder level in four minutes. It is evident at once that a large amount of ability must be present before such a test can begin to measure, and that speed is practically the only element of ability measured. Since speed is desirable, however, it is highly worth while to measure it, although we must be careful not to take a measurement of speed as a sufficient index of ability. S6 THE FIFTEENTH YEARBOOK In arranging the language scales, the writer has assumed that older children will not only be able to do the same tasks more rapidly and more perfectly than younger children, but that the older children will also be able to accomplish perfectly tasks which the younger children coiild not begin to do. The results from twelve or thirteen thousand children, upon whom these sentences have been tried, seem to support this assumption in every respect. The unit of difficulty used in constructing Language Scales B, C, D, and E is the P.E., or median deviation from the median of a school- grade distribution of ability, assuming the curve of distribution for ability in any given grade to be "normal" in shape and equivalent in range to the range of any other grade distribution. The reasons for making these assumptions and using this measure of the variabiUty of a 5cnfertce Ho. 1 1 ® G © S 0 11 © 23 © 31 35 58 © © © 4S © 54 © P.E. Scale I 3 4 i L ' ' ' 6 7 8 9 10 11 Fig. 2. — ^Linear projection of Language Scale B. grade as the unit need not be discussed here. It is worth while to remem- ber, however, that if we take any two groups, selected by the same or by equally capable judges as representing two different degrees of the same sort of ability, these two groups are very likely to be distributed "nor- mally" in each case and the variability of one group will probably be very nearly equal to the variability of the other. In other words, approximately as many errors of judgment are made in selecting one group as in selecting the other, and in each group approximately as many are overestimated as are imderestimated, errors in either direction being much more frequently small than large. In brief, the P.E. is a convenient vmit which has approximately the same value in every grade and may therefore be used to measure the distance between grades and the difficulty of tests for all grades. On a P.E. scale above an arbitrary zero point, the location of each sentence of Language Scale B is as represented in Fig. 2. It will be observed that the intervals between sentences are not exactly i P.E. in each case, but for practical purposes they are near enough. The improve- ment made by a person who was at first able to complete only two sentences but is now able to complete four is at least in a very real sense COMPLETION TESTS FOR PUBLIC-SCHOOL USE 57 equal to the improvement made by the person who was at first able to complete only seven sentences but is now able to complete nine. As yet the writer has not had time to use the new scales very exten- sively in their present form, although the sentences of which they are composed have been thoroughly tested in other combinations of sen- tences. Language Scale A/ which is so poorly graded as not to be worthy of the name "scale," has nevertheless revealed some of the possible values of the completion test in school procedure. Scale A was composed of 24 sentences, so that with 2 points credit for each perfectly completed sentence the maximum score possible would be 48 points. The median score and the range of the middle 50 per cent of the scores are shown in Table I for those pupils of the elementary grades who were tested with Scale A. TABLE I Scores of Elementary-Grade Ptxpels on Language Scale A School Grade II III IV V VI \ai VIII No. tested Median score. . . . 25 percentile. . . . 75 percentile .... 1318 4-59 2.40 6.57 1437 8.99 6.22 13 04 1463 1433 10.28 18.60 1507 18 -39 1497 22.02 1454 21.92 18.14 25-36 1456 25-27 21.58 29.08 1427 28.06 24-34 32.42 In connection with Table I it is worth while to consider Table II, which shows the median scores of twelve-year-old boys according to the school grade in which they are found. TABLE II Medlan Scores of Twelve-Year-Old Boys on Language Scale A School Grade III IV V VI VII VIII No. tested 22 IO-5 62 12.9 155 16.3 269 21.6 217 24.7 76 29.8 Median score ' M. R. Trabue, "Some Results of a Graded Series of Completion Tests," School and Society (April 10, 1915), 537-40. 58 THE FIFTEENTH YEARBOOK It would seem from Tables I and II that the ability measured by these completion-test sentences is related rather closely to the ability which teachers consider when promoting pupils to higher grades. Table III, showing the median scores of sixth-grade boys according to their TABLE III Median Scores of Sixth-Grade Boys on Language Scale A Age lO II 12 13 14 IS 16 No. tested 70 234 201 22.6 243 21.9 164 21. 1 74- 18.9 30 17.6 7 133 Median score ages, is interesting in this connection. This table furnishes evidence of a fact which needs to be emphasized. Teachers tend to retard the bright young chaps because they are "too young" and to promote the dull old fellows because they are "too old" for the grade. Possibly the most serious retardation problems in our American school systems arise from our lack of adequate provision for the exceptionally bright young children. Scale A was given in a number of intermediate schools with interest- ing results. At Decatur, Illinois, four classes of Grade VIIA were tested. The median scores were as follows: Class VIIA' VIIA" VIIA3 VIIA< No. pupils. . . . Median 23 28.3 22 26.9 20 25-5 21 23-7 After this result was noticed, I was interested to find that each pupil is classified, on entering VIIA, according to the judgment of previous teachers as to his general ability to do school work, those of greatest ability being put in VIIA% the next best group in VIIA% and so on. In Grand Rapids, Michigan, those pupils in the eighth grade whose superior scholarship has been proved are allowed to take up the study of Latin. The median score of the 28 pupils in the Latin group was 34.0, with only one score as low as 27, while the median score of the 71 other COMPLETION TESTS FOR PUBLIC-SCHOOL USE 59 pupils in the same grade of that school was 29.8, with two scores of 17. Evidently the ability required to make high scores on Language Scale A is rather closely related to the quality for which teachers look when they attempt to select their brightest pupils for special work. In view of the fact that ability to complete incomplete sentences is so obviously dependent upon abiUty to read and interpret printed words, it is not surprising to find that home training and nationality show an appreciable correlation with ability in these language scales. Three graduating classes in the elementary schools of Kansas City were measured on Language Scale A, with the following results: School Hyde Park Whittier Hamilton No. pupils Median 44 33-4 42 31-3 29 27-3 The Hyde Park and Whittier schools are distinctly American, the first being in the best residence section of the city and the second being in a respectable middle-class district. The Hamilton School, however, is in a foreign district, one block from the Swope Settlement. The writer found a similar situation in Bayonne, New Jersey. Three VIA classes were measured on Scale A, with the following results: School Mann Vroom Lincoln No. pupils Median 32 23.6 37 20.4 39 19.2 The Mann School is largely American, the Vroom largely Jewish, and the Lincoln largely Slavonic in student population. The completion test is not proposed as a substitute for the judgment of teachers in promoting pupils, but rather as an aid to it. Employing one of these scales will call attention to those children who have unusual ability to understand and interpret printed words and phrases. These unusual cases may then be investigated more carefully and such read- justments made as will allow these children to work under the most favorable conditions. PART II APPLICATION OF SCALES AND UNITS OF MEASUREMENT IN EDU- CATIONAL SUPERVISION AND ADMINISTRATION CHAPTER V WORK OF THE DEPARTMENT OF EDUCATIONAL INVESTIGATION AND MEASUREMENT, BOSTON, MASSACHUSETTS FRANK W. BALLOU Department of Educational Investigation and Measurement, Boston, Massachusetts The three phases of the work of the Department of Educational Investigation and Measurement, which will be described in order below, are: (A) measurement of educational results; (B) supervision of a revision of the elementary course of study; (C) the organization of a plan for the promotion of teachers on merit. A. MEASUREMENT OF EDUCATIONAL RESULTS I. Arithmetic. — The Courtis standard research tests in the four fundamentals were first given in Boston in October, 191 2, by Mr. Courtis himself. Later they were given by others and have been con- tinued to be given by this department. Their introduction into the Boston schools has been gradual, beginning with 21 districts in October, 1912, and covering the 70 elementary districts in May, 1915. The tests have resulted as follows : 1. They have made possible the establishment of objective stand- ards of achievement for Grades IV- VIII in addition, subtraction, multi- plication, and division. These standards are based on the median score attained by large groups of pupils, and represent, therefore, the minimal achievement of at least 50 per cent of the children tested. These stand- ards are shown in Table I. 2. These tests have shown the relative standing of each school, of each class, and of each pupil, in the 70 elementary school districts 61 62 THE FIFTEENTH YEARBOOK tested, thereby providing the administrative officers with information as to the conditions which need improving. 3. They have revealed the ineffectiveness of the present general class drill in arithmetic on the four fundamentals, by showing that approxi- mately one-third of the class gets more drill than it needs, another third makes fair progress, and the other third not only does not improve but, in many cases, actually loses in ability. TABLE I Time Aixowed Grades Addition, 8 Minutes Subtraction, 4 Minutes Multiplication, 6 Minutes Division, 8 Minutes Attempts Rights Attempts Rights Attempts Rights Attempts Rights IV 8 9 10 11 12 6 7 8 9 II 8 9 10 II 12 6 7 8 9 II 6 7 9 10 II 4 5 7 8 10 4 6 8 10 12 2 V 4 6 VI vn 8 vin II 4. They have demonstrated the need of drilling pupils in those four fundamentals in which they are deficient. To this end, a large number of the 70 elementary school districts have introduced several kinds of practice material in arithmetic, the relative merits of which are being studied by teachers and masters with a view to using that which proves most effective. II. Spelling. — Proceeding from the fact that an eighth-grade pupil probably uses not more than twenty-five hundred words in his writing, whereas the spellers in common use contain from ten to fifteen thousand words, the department has worked as follows: 1. With the aid of many teachers minimal and supplementary lists of alleged difficult words have been prepared for each of the eight grades. These lists were placed in the hands of each teacher in the elementary schools, with the suggestion that the minimal list be emphasized in the spelling instruction {School Document No. 8, 19 14). 2. The intention is by no means to limit the instruction to the words which are at present contained in the minimal list, but it is to make sure that the child learns to spell the words which he actually uses in his voluntary writing. EDUCATIONAL INVESTIGATION AND MEASUREMENT 63 3. A test was given in May, 191 5, to all grades above the second, for the purpose of determining the relative degree of difficulty of the words in the minimal lists. 4. As a resvdt of that test each word is accompanied by a percent- age, which indicates the munber of pupils that spelled the word correctly. In this way each teacher is furnished with a list of alleged difficult words, together with information as to their relative degree of difficulty, thus making it possible for a teacher to place the emphasis in her instruc- tion on the more difficult words {School Document, No. 10, 1915). 5. A further result of this test is to fiurnish each teacher with a standard by which she may judge whether her class is above or below the general standard for the city. If, for instance, a word is indicated as having been spelled correctly by 90 per cent of the children of the city, a teacher knows that if more than four out of the forty pupils in her class misspelled that word her class is below the standard ability of Boston children to spell that word. 6. A study has been made of the various lists of words that have recently been prepared as a result of scientific investigation, and the resulting list has been supplied to the teachers. It is the ultimate intention to make selections of words from this list to be added to the minimal lists until the minimal lists contain practically all of the words which are within the writing vocabularies of the normal pupils of each grade. III. English. — Before the Department of Educational Investigation and Measurement was organized a Committee on Standards in English had been at work for some time and had set up some tentative require- ments in English which the committee felt ought to be met by every pupil who graduates from the elementary schools. These requirements were approved by the Board of Superintendents and thus became authoritative standards. The requirements are as follows: Every graduate should be able: 1. To copy twelve hnes of simple prose or poetry, and a bill of at least seven items. (Copying is not an end in itself, but a means to an end. The pupil should be made to see that accuracy in arithmetic, language, and other subjects may depend largely on accuracy in copymg.) 2. To take down from dictation a passage of simple prose. (The purpose of dictation is to test language forms, punctuation, and spelling 64 THE FIFTEENTH YEARBOOK already taught. It should never be used as a method of teaching. It should succeed and not precede a teaching lesson.) 3. To write from simple directions a friendly letter or an applica- tion for a position. (The letter is to be the pupil's own work, but he may be allowed to make corrections and to re-write. There should be no corrections by the teacher.) 4. To write within a half-hour a simple, original composition of not less than one page of letter paper, with every sentence grammatically complete. The pupil may make revisions, including interlinear cor- rections, but must not re-write. In this composition the total number of serious errors in grammar, spelling, and punctuation should not exceed five — such errors, for example, as ''I seen," "we was," "had wrote," "he try" for "he tried," "a women," the use of "they" for "there," "there" for "their," "to" for "too"; the misspelling of such common words as "Wednesday," "February," "eighth," "which," "stopped," "nineteen," "minute," "father," "mother," "English"; the omission of the period at the end of a sentence. 5. To recognize the parts of speech in their common uses; to explain the construction of words and phrases in a simple sentence containing not more than one phrase modifier in the subject and one phrase modifier in the predicate; to have a practical understanding of the uses to which the dependent clause of a complex sentence can be put — whether it be to serve as noun, adjective, or adverb; to know the principal parts of regular verbs and of the common irregular verbs, and their tense forms through the indicative mood. 6. To read at sight with readiness and good expression simple prose as difficult as Little Men or Hans Brinker. 7. To quote either orally or in writing fifty lines, not necessarily consecutive, of classic prose or poetry. (The pupil should look upon this not merely as something to be expected of him in the high school but also as a part of his equipment for life.) 8. To stand before the class and talk clearly on some subject of personal, school, or public interest. The Committee on Standards in English is co-operating with the department in putting these requirements into effect. To this end, the department and the committee caused two tests to be given in EDUCATIONAL INVESTIGATION AND MEASUREMENT 65 November, 19 14, to 4,944 pupils in the first-year classes in the high schools. These tests were in accurate cop3dng and in written memory work. 1. The test in accurate copying was to discover what degree of accuracy should be expected of pupils when they are asked to copy fifteen lines of prose in fifteen minutes. 2. The test in written memory work was to find out how well pupils remember the fifty lines of classic prose and poetry which the course of study requires that they shall have committed to memory before grad- uating from the elementary schools. 3. As a result of the test in accurate copying it has been found that boys will copy fifteen and one-half lines (4I inches long) of prose in fifteen minutes, with 50 per cent of the pupUs making less than five errors of any kind (spelling, capitalization, punctuation, words oniitted, words added, wrong words used, misspelled words, undotted i's and uncrossed t's). 4. The same test shows that the girls will copy more than sixteen lines in fifteen minutes, and that 50 per cent of the girls will make less than three errors of any kind. 5. By giving a standard test under controlled conditions, it will be possible to give a similar test at some future date and determine whether or not improvement has been made. IV. Geography. — In co-operation with the head of the Department of Geography in the normal school the department has given two tests to pupils in the eighth grade to determine: 1. How well pupils in the eighth grade remember the geography which they were taught in the earlier grades. 2. What abiHty pupils have to reason about geographical data. 3 . Whether mere place-geography is being overemphasized in teaching. V. Penmanship. — The quality of the handwriting of elementary- school graduates has been studied as follows: 1. Six-hundred specimen papers were selected from the 4,944 papers written in the test in accurate copying. The pupils were not aware that their handwriting was to be examined, hence they probably wrote in a natural, unrestrained manner. 2. These 600 specimen papers were rated according to the Ayres' Scale for Adult Handwriting, using only the specimens in the scale 66 THE FIFTEENTH YEARBOOK under 90, 70, 50, and 30. Specimens poorer than the 30 specimen in the scale were rated 10. . 3. The specimens were rated by a committee of six teachers who are superior teachers of handwriting. Each paper was rated by three persons. 4. The final rating of a paper was determined in the following manner: Where two or more persons agreed, that rating was given the specimen; where no two agreed, the middle rating was assigned. 5. A further study is being made of the merits and defects exhibited in these 600 specimens and a report will be printed and distributed among the teachers. B. REVISION OF THE ELEMENTARY COURSE OF STUDY In co-operation with two of the assistant superintendents the depart- ment is supervising the revision of the course of study in the elementary schools. In order to assist the teacher to economize her time and energy by the adoption of more definite purposes in teaching, these special features are being introduced into the course. 1. A concise, definite statement of the aims to be accomplished in the teaching of each subject in each grade. 2. A statement of the irreducible minimal essentials in each subject in each grade. 3. A definition of the objective standards of achievement in various subjects as far as they have been worked out. This revision is being made with the co-operation of about 40 com- mittees, including 359 different teachers, working on the following sub- jects: arithmetic, reading, stories and literature, spelling, grammar, composition, dictation, geography, and history. This utilization of the knowledge, ability, and experience of the teachers will be followed by further professional educational advice from principals and superintendents. This method of course of study revision has the advantage of building up a practical course of study based on classroom experience, of securing the sympathetic understanding by the teacher of the course when it is adopted, and of affording helpful stimulus and proper encouragement to the teaching staff which must follow from such professional recognition. EDUCATIONAL INVESTIGATION AND MEASUREMENT 67 C. ORGANIZATION OF A PLAN FOR PROMOTION OF TEACHERS ON MERIT Inasmuch as the higher positions in the school service must be filled by the superintendent when vacancies occur, the department has pro- ceeded on the assumption that any plan of promotion, honestly admin- istered, is better than no plan at all. The need of a systematic plan is seen in the following phases of the conditions in the Boston school system. 1. The size of the public-school system, with its nearly two thou- sand elementary and over five hundred high-school teachers, makes it impossible for the superintendent to know the work of the teachers except indirectly. 2. The variety of ranks both in the elementary and in the high schools makes relatively a large number of promotions within the service. 3. The large number of candidates who hold certificates making them eligible for promotion makes necessary some further plan for determining their relative professional qualifications. The scope and method of the work of the department to formulate a plan will be illustrated by the following brief statements: 1. Several conferences have been held with sub-masters and master's assistants, because they are among those most interested in a plan of promotion on merit. Among other things, it was agreed, subject to the necessary official approval, that (a) the sub-master in the school where the vacancy occurs should be given first consideration for appointment; (b) other qualifications being equal, that sub-master in the service longest as sub-master should be appointed first; (c) ratings of teaching ability and estimates of probable future success should be secured from (i) assistant superintendents, (ii) masters of schools, (iii) director of promotion and educational measurement. 2. At a conference with the Board of Superintendents a common basis for securing discriminating and comparable ratings was agreed on. 3. A comprehensive basis for judging merit has been prepared by the department after a study of the available plans of rating teachers in cities throughout the country (see forms 264, 265, 266, and 267). 4. From April, 19 14, when the department began work, to June I, 191 5, the following higher positions have been filled in accordance with the proposed plan, in so far as it has been worked out at the present 68 THE FIFTEENTH YEARBOOK time. All appointments to higher positions between these dates have been made according to the merit system. In elementary schools: First assistant, grammar Women i First assistant in charge Women 7 Master's assistant Women 5 Sub-master Men 5 /Men S Master < ,,r 1^ Women i In high schools: Master, head of department Men i Head master Men 4 CHAPTER VI THE APPLICATION OF STANDARD MEASUREMENTS TO SCHOOL ADMINISTRATION D. C. BLISS Superintendent of Schools, Montclair, New Jersey A study of the educational literature of recent years reveals a steadily increasing interest in the possibility of the application of standard tests and measurements to the problems of school administration. Many prominent educators honestly question the value of such tests, because they feel that the most important elements in the mental and moral development of pupils are of an intangible character which can by no possibility be confined in terms of measurement. No one would dispute the fact that human life is a deeper and more complicated subject than can be probed by quantitative tests; nevertheless, when the more subtle components have been excluded, there remain some essential elements in education which are purely objective, and that these can be measured with reasonable exactness there is no reason to doubt. Because some things of supreme importance cannot be included in this category is no valid argument for rejecting the entire plan. We measure a man in terms of achievement; to apply to the schools the same test, the ability to produce results, is only logical and reasonable. The spiritual side of education is real and in all probability defies measurement, but a com- plete education includes elements other than the spiritual, and so far as they are present they can be measured. If they are so vague and indefinite as to escape measurement, their existence may well be doubted. The more clearly the objective results of education are understood, the greater the appreciation of the spiritual elements in the child's training. To reassure those who feel that any effort to arrive at a definite knowl- edge of educational values carries with it the danger of obtruding com- mercial methods into the region of things of the spirit, the well-proved truth may be reiterated here: the more clearly the objective results of education are understood the greater is the appreciation of the fact that unless all three natures, physical, mental, and spiritual, are being 69 70 THE FIFTEENTH YEARBOOK definitely led toward their fulfilment, no system of child-training is even approaching the adequate performance of its function. The mechanical application of standard tests with the resultant accumulation of medians, graphs, and charts is in itself a futile thing. Only as these tests are informed and controlled by a trained and sym- pathetic mind using the facts revealed as the basis of a constructive policy for future work do they find their justification. They show to a superintendent the extent to which his plans have been correctly inter- preted and put into operation, and they furnish him with a sound basis for necessary changes, whether in the way of revision or of the introduc- tion of new methods. Because of the increasingly heavy demands upon the public schools, economy of time and of energy in every direction has become a necessity. Consider the relatively simple subject of penmanship. Once a pupil's handwriting was acceptable if the letters were well formed and the lines even, and the method of achieving this result was left largely to nature, or, when she had failed in the bestowal of the necessary gift, to the child's dogged patience in drawing over and over something resembling the copperplate sentence at the top of his writing-book. Now, writing has been differentiated from drawing, and every motion has been ana- lyzed; unaccustomed muscles must be trained and co-ordination estab- lished; form must still be maintained, but the emphasis has been shifted to power; legibility must be accompanied by speed and a degree of freedom which makes it possible to continue the rapid, even movement for long stretches of time without fatigue. The problem before the school is how to meet these entirely reasonable demands without unduly prolonging the amount of time which may fairly be assigned to the subject. To this end the department of superintendence must know unmistakably every point of success or failure in the penmanship teach- ing and drill throughout the whole school system, that waste of time and effort may be eliminated. By what method shall the facts be ascer- tained most quickly and effectively ? Spelling is another comparatively simple subject, with a compara- tively short time allotted to it in the school program. Questions like these face the teacher and the superintendent; Is the method in use producing accuracy ? Is drill being wasted upon unusual words which will in all probabiHty never find a permanent place in the child's vocabu- lary ? Is drill on many words being carried beyond the point necessary STANDARD MEASUREMENTS AND SCHOOL ADMINISTRATION 7 1 for fixing them in the memory? Has every child a limited working vocabulary which it is impossible for him to spell incorrectly ? Unsupported opinion can no longer be brought forward to decide, even in cumulative fashion, these and manifold other points which arise. What twenty eminent educators think as to the efficacy of a certain method gives way before the facts which careful investigation shows. Valid conclusions can be based only upon the results of accurate tests applied to large groups and continued for a sufficient length of time to eliminate possible error. If, as has already been said, such tests result only in the mere accumulation of educational statistics, they fail to justify the attendant expenditure of time and energy. They must surely function in an improvement in schoolroom practice or they are worse than useless. Unfortunately, it is precisely at this essential point that the vitality which should inform the laboriously acquired statistics is allowed to escape, leaving an inert mass of figures to show that someone has been busily and vaguely occupied about something. In education, as in every other department of life, the demand is not "Be busy about something," but "Be busy to some purpose." So simple material as the relationship of "age and grade" should be carefully interpreted, and the inferences drawn from it should be a factor in shaping the school policies for the ensuing year. Percent- age of promotion, "mortality" both in the grades and in the high school, the number of pupils accelerated or retarded, and other data of this character furnish evidence for a diagnosis of the health of the school body. As a case in point may be cited the age-and-grade table of the Montclair, New Jersey, schools for September, 19 12, which showed 23 per cent of retardation. Since this seemed too large for such a com- munity, the superintendent and principals met for a discussion of plans looking toward the improvement of the situation. No radical steps were taken, but the schools were fully aroused to their responsibility for the excessive number of repeaters, with the result that the needs of individual children were given more careful consideration and a greater degree of flexibility was infused into school administration. In Sep- tember of the following year figures were compiled on the same basis and the totals, which showed a decided improfvement, formed the subject of another discussion. Four years of this policy have reduced the percentage of retardation almost 50 per cent. 23.39S Chart I.— Percentage of retardation in Montclair schools in successive years, showing steady reduction. STANDARD MEASUREMENTS AND SCHOOL ADMINISTRATION 73 The same general plan for determining values can be applied to such questions as the wisdom of the establishment of open-window classes, which, it may be said in passing, should be distinguished from the open- air classes — a wholly different problem. The argument in favor of open- window classes seems faultless. Fresh air is essential to health; it has proved effective as a healing agent in cases of tuberculosis. Anaemic children are benefited by open-air rooms. The inference is that healthy children will show a still greater improvement, owing to their better physical status. The theory seems flawless, but we need to assure our- selves that the facts accord with the theory, and the only way to deter- mine this is to test it. One test is to weigh the children of the open- window class at certain intervals and then compare the change in weights with the change in weights of an equal number of children who are in a classroom of the usual type. The chances are that a superintendent who has before him such a record as is shown in Chart II will hesitate before extending the plan to his entire school system. Especially will this be true if other tests of a different character point to the same conclusion. In dealing with subnormal pupils the opinion generally prevails that they should be given a form of education in which the manual arts pre- dominate. It is also believed that defective children who seem to learn readily and are able to recite fluently are doing it in parrot-like fashion and that even this seeming facility will not long persist, that it is simply a case of learning something today and forgetting it tomorrow. Obvi- ously, it is of great importance to the teacher to know whether this assumption is true of the individual pupils in her care, or whether she may stress the academic phase of her work with some hope of making an enduring impression. Only by testing the individual pupils is she able to determine the character of the training suited to each. The possibilities of the standard scale for such children are shown by the records in penmanship and arithmetic of a subnormal class last year. A local scale, based upon that of Dr. Thorndike, was used to measure the quality of the handwriting of all members of the class, and this formal rating was made once in two months. The arithmetic scale consisted of ten problems in fundamental operations. A definite time- limit was fixed for the solution of these problems. No credit was given for incorrect results or omitted examples. The records made by a sub- normal pupil in the two subjects appear in Charts III and IV. 74 THE FIFTEENTH YEARBOOK •3fr •SO- 39^ A\-\Z •m HOH 6-4 1-9 3-6 sisi &I4 16-1 15-15 1442 4*^ i i I I MWtoDEC DEC to FEB- FEB-tbMARCH MARrtoMAY Chart II. — Gains and losses in weight of open-window class and of control class at Montclair, New Jersey. (The first column in each time group represents the open- window class, the second the control class. The entire height of the columns shows weight gain while the white portion records weight loss. Thus, from November to December the pupils of the open-window class gained 23 lbs. 5 oz. and lost 6 lbs. 4 oz., making a net gain of 17 lbs. i oz., which is shown by the blackening of the upper portion of the columns. Diuring the same period the pupils of the control class gained 23 lbs. 9 oz., and lost i lb. 9 oz., leaving a net gain of 22 lbs. In every time group the control class showed a greater net gain than the open-window class.) STANDARD MEASUREMENTS AND SCHOOL ADMINISTRATION 75 7W/ Sept 8 Jain. ^ 10 Mar. V/ ^ II JUNE^ i2l \z Chart III. — Performance in handwriting of a subnormal pupil. (The quality of the handwriting is indicated by the length of the hatched oblong and also by the digit at the right. The improvement is steady from September to June.) Chart IV. — Performance in arithmetic of a subnormal pupil. (The figure at the right shows the percentage of correct answers.) 76 THE FIFTEENTH YEARBOOK In the purely mechanical process of writing there would seem to be no doubt that this pupil benefited directly from the teaching. In arith- metic the progress is not so evident. The fact that the arithmetic test was confined to examples in funda- mental operations of the same degree of difficulty made the tests prac- tically uniform. It is very doubtful if a pupil whose record fluctuates to the extent indicated above receives any permanent value from such exercises. Similar charts made of every pupil in the class showed the same general tendency; improvement in penmanship was fairly constant, while in arithmetic frequent lapses appeared. By the use of the standard scale the superintendent is enabled to check the results of any experiment in his schools and banish, from his own mind at least, any lingering doubts as to the wisdom or lack of wis- dom of what he has undertaken. An experiment in one of the Montclair schools with a precocious class is a case in point. In September, 191 2, in this school, a group of fourth-grade children of fairly uniform and TABLE I Record of Tests Fundamen- Fractions English Spelling Writing tal Operations 83.6 81 98 II. 6 78 86.9 80 98 II. 2 80 77 71 97 II. 4 69 87 83 97 II . I 83 Superin- tendent's Test Com- position Watchung VIIA. . WatchungVIIB.. Grove VIIA Special group . . . . 42.2 45-4 45-4 45-8 superior ability was put in charge of a strong teacher who was instructed to allow the class to advance as rapidly as it desired. No pressure was ever brought to bear upon the pupils, but dawdling was discouraged. The class remained with the same teacher for two years and in this time did three years of work. Four months after the special group entered the seventh grade, careful tests were made to determine to what extent the experiment had been a success or failure. A comparison of the record of the special seventh-grade pupils with that of the entire seventh grade with which they had been merged and with two other seventh grades of a similar type showed a very gratifying situation. Tests in spelling, arithmetic, and EngHsh were given by the principal and a STANDARD MEASUREMENTS AND SCHOOL ADMINISTRATION 77 standard test in composition by the superintendent, while the penman- ship was rated by the writing supervisor. A conclusion based upon the results obtained from these three separate and independent sources can fairly be presumed to represent the facts (Table I). V mi 1915 1915 W////////////////////////////M. ^^^^^^ VII VIII 8.6 \Z 8.6 12.7 5.2 13.8 5.2 14.3 Chart V. — Improvement in penmanship in Grades V-VIII at Montclair, New Jersey, resulting from the use of a standard measuring scale. (The upper portion of each horizontal band represents the quality of the penmanship in the grade indicated in September, 1912. The lower portion shows the quaHty reached in Jime, 1915.) It should be noted that the records in spelling, fundamental opera- tions, complex fractions, and English show rank in percentage, while the penmanship rating and the superintendent's test are given in points and should be interpreted only as furnishing a basis for comparison. That the children had not been advanced too rapidly is further indicated by the fact that all of them were regularly promoted at the 78 THE FIFTEENTH YEARBOOK end of the year to the eighth grade, where they are now doing excel- lent work. Standard tests are not only useful in bringing out the facts concern- ing pupils and classes, but are equally valuable in indicating general tendencies in the system as a whole. A recent school survey revealed the fact that in a certain system of schools the pupils in the fourth grade wrote practically as well as those in the fifth grade, while the hand writing of sixth-grade pupils was actually better than that of children in the seventh grade. Such a departure from a normal curve of advance may be justified by some unusual condition. However, the supervisory department should be fully aware of the facts if effective remedial measures are to be applied, and no amount of theorizing will furnish these facts. They can be obtained in only one way — by the application of one of the standard tests in penmanship. With the returns from such a test tabulated and charted, the situation in every grade is at least reasonably clear and a definite policy can then be formulated with some chance of reaching the root of the dijficulty. The effect of the use for three years of a standard penmanship scale upon the quality of the writing of public-school children is indicated in Chart V. It is perhaps impossible to show by figures the total effect which tests of the type indicated here produce in the quality of the work of the schools. When purposeful effort is substituted for aimless drift- ing, there can be no permanent withholding of successful results. The only limits to investigation of this kind are those determined by the time which the supervisory department can afford to give to the work. It is based upon the fundamental idea of a continual local survey, made by those who know the actual school conditions, and who are seeking the facts as they exist, for the sole purpose of formulating constructive policies. CHAPTER VII A HALF-YEAR'S PROGRESS IN THE ACHIEVEMENT OF ONE SCHOOL SYSTEM H. G. CHILDS Associate Professor of Education, Indiana University SECTION A. THE PROGRESS AS MEASURED BY THE THORNDIKE VISUAL VOCABULARY TEST The following report is based upon the responses of 754 pupils in the Bloomington, Indiana, schools to the Thorndike visual vocabulary test, which was given during the first week in February and again the first week in June, 1915. Only the papers of those pupils were considered who wrote both tests. The grades included are the IVB to the VIIIA inclusive. The usual method of tabulating the results has not been followed, as the writer thinks that method not the best suited to show the actual achievement of any pupil; for example, the first two papers the writer took from a pile were each scored "line 7." In the first of these papers the pupil had made the correct response to every word in lines 4, 5, 6, and 7 and to one word of line 9. In the other paper correct responses had been made to every word in lines 4, 5, and 6, to four words of line 7, to two words each of lines 9 and 10^, to three words each of lines 8 and 10, and to one word of line 11. While both these papers were scored as of equal value, the writer is convinced that the ability of the second pupil should be rated as nearly double that of the first. By the method of scoring used in this report, the first pupil receives a mark of 119 and the second a mark of 214.3. Assuming that the line values of the scale are correctly assigned (the Indiana results indicate that radical revision is needed), and that each word within any line is approximately of the same value as any other word of that line, why not assign a definite score for each word depend- ing upon the line it is in? Such a marking scheme will give a pupil 79 8o THE FIFTEENTH YEARBOOK credit for what he actually accomplishes in the test, and will have the added advantage of stating the pupil or class ability in one numerical result which can be treated statistically with greater ease than in the Thorndike scheme. The papers in this report have been graded on this basis: every correct response in line 4, four points; line 5, five points; line 6, six points, etc. — line eleven, 18 . 3 points. The sum of all the word points represents the total value of the paper. For comparative purposes the score of any paper or class median can be reduced to line and word values of the scale thus : line 4 has a range of value from o to 20; line 5, 21-45; ^^^ 6, 46-75; line 7, 76-110; line 8, 111-150; line 9, 151-195; line 10, 196-245; line io|, 246-297.5; and line 11, 298-352.5. By this scheme the ability of a pupil or of a class may be stated in fractional parts of a line or in terms of the first, second, etc., words within a given line; e.g., a pupil whose score is 126 has an ability equal to that represented by line 7 and 16 points additional, which is four- tenths of line 8, or two words of this line. Table I shows the results of the February and June tests. The features to which attention is called are: (i) the score of each grade in the February test, (2) the gain from grade to grade in the February test, (3) the gain of each grade from February to June. From this table the median gain from grade to grade by the Feb- ruary test is 18.8, or, stated in other terms, this gain is the equivalent of three words of line 6 or of two words of hne 9. The greatest difierence, 38.5, is between Grades VIB and VIA; and the least, 6.7, is between Grades VIIB and VIIA. The median gains within each of the grades between the February and June tests is 17.3; the highest, 37.5, is in Grade VIB and the lowest, 3 . 9, in Grade VIIIA. It will be noted that the curves of Chart I show no marked plateaus and that the achievement in the VIB grade is double that in the IVB, while the VEIIA median is three times that of the IVB. This seems to indicate a steady and normal vocabulary development. As comparable tests were not available, the same tests were given in June as in February. Whether the gain during the term is due to the natural growth in vocabulary on the part of the pupils or to their having remembered some of the words from the February test and looked A HALF-YEAR'S PROGRESS IN ONE SCHOOL SYSTEM 8i them up afterward as a matter of curiosity, I camiot say. I think the latter is true in only a very few cases. TABLE I Thorndike Visual Vocabulary Test (754 Cases) Grade No. of Pupik Date Score P.E. Line Value February Difference between Grades Gain during Term IVB .... 91 85 65 102 81 83 62 69 60 56 Feb. June Feb. June Feb. June Feb. June Feb. June Feb. June Feb. June Feb. June Feb. June Feb. June 81.3 92-S 107.5 134-8 129.0 151. 0 145-3 156.0 164. 1 201.6 202.6 219-5 212.8 229.8 219-5 240.3 238.6 256.1 254-1 258.0 28.S 28. 5 34.8 35-6 48.8 41.6 34 -o 30.4 43 0 43-5 28.0 34-1 33-0 31-7 35-0 37-3 35-0 32-9 36.7 48.3 7.18 7-S 7-93 8.62 8.47 9.02 8.88 9-13 9-31 10.12 10.15 10.45 10.36 10.7 10.45 10.91 10.87 ioj.21 10I.17 10I.24 IVA . . . . VB VA VIB .... VIA .... VIIB ... VIIA ... vmB .. ViJIA .. 26.2 2I-S 16.3 18.8 38. 5 10.2 6.7 19. 1 iS-5 II. 2 27-3 22.0 10.7 37.5 16. g 17.0 20.8 17-5 3-9 Grade median Feb. Jixne 183.3 210.6 34-9 34-8 9-73 10.28 18.8 17-3 To determine if any correlation existed between estimated teaching efficiency and class improvement in the vocabulary test, the writer secured the ratings of several competent judges on a considerable group of teachers in the grades from the IVB to VIIIA. In the IV and V grades where the work was non-departmental, the coefficient of corre- lation was negative ; in the VT, VII and VIII grades where the teachers 82 THE FIFTEENTH YEARBOOK Grade 4B 4A sB sA 6B 6A 7B 7A 8B 8A Line Values 260 240 220 180 160 140 80 60 ■T ""T"!"!"""! " " " " " t t"-' 1 1 ~ rf* ^ -- _ -i^ -. ^ - 24.'; ^-^ ^^ : . ^^ -.---;j- .^. ........ ^^ / - - ^** -I ^- 4- .__^6; ,^:„ "5 Z^^ * ' »*'' y - ^ ^ -'--•- "t"\ .-^ . j^j ~ ~ -7 "7 ■" " Line 0 J / : : ; ! — -. -J I 7 . ■ ^ r . :" : i^''z."'. -..^-'--/- . _ . ISO V — y --_ : ..2 ^t -; "^ : ^^ ^^ ^ i J J J ...... ..t~~.J~.~^~. ---. :::::::^:::z_::: ::::::::::: no r / . "~ t.J. - , 1 Luic '/ 3"?" — t t _ A^ -% - f ............................ ^^ 1 _ Line 6 L _ -a -^— February June Chart I. — Graphic representation of the progress in visual vocabulary of 754 pupils at Bloomington, Indiana, based on Table I. A HALF-YEAR'S PROGRESS IN ONE SCHOOL SYSTEM 83 of reading and literature were employed for their ability in this line, the coeflScient was between 0.6 and 0.7. SECTION B. THE PROGRESS AS MEASURED BY THE COURTIS TESTS, SERIES B During the second semester of the school year 19 14-15, the writer, in conjunction with Mr. H. L. Smith, then superintendent of the Bloom- ington, Indiana, schools, gave the Courtis arithmetic tests in all grades of the elementary school from the IVB to the VIIIA, inclusive, for the purpose of measuring the growth in arithmetical achievement of pupils in this system. The tests were given during the first week in February and again the first week in June. The records of the 809 pupils who wrote both tests alone are included. Table II indicates the number of pupils in each grade who wrote both tests and also the median number of attempts and rights, the per- centage of accuracy, the percentage gain in rights, and the percentage gain in accuracy, for each of the four fimdamental processes and for both the February and the June tests. From this table the gain in achievement from grade to grade may be noted for the same date and for each of the processes; also the growth in achievement for each process between the February and June tests for each grade; and relative achievements in the different processes for the same grade and date may be compared. In explanation of the results here set forth in comparison with those of other systems, it may be noted that formal work in arithmetic is begun in the Bloomington schools in the IIIB grade. Table II and the accompanying charts show the following features worthy of note: (i) the slight gain from grade to grade in the February addition results for rights; (2) the marked gain in rights in addition of the June over the February results in Grades VI, VII, VIII; (3) a marked gain in rights in subtraction from grade to grade, but a decreasing ratio of gain between the February and June tests from Grade IVB to Grade VIIIA — there being an actual loss in achievement in rights in Grades VIIA, VIIIB, and VIIIA; (4) the shght progress in achievement in multipUcation from grade to grade and the lack of improvement of the Jvme over the February results except in Grade IV; (5) the marked gain in division from grade to grade in both tests but the lack of any marked improvement of the June over the February results in any grade — there being a moderate gain in Grades IV, V, and VIIIA; (6) that the actual achievements in attempts, rights, and accuracy in Grade IVB 84 THE FIFTEENTH YEARBOOK TABLE II ADDITION Percent Gain in Grade No. OF Date Attempts Rights AcCtJRACY Rights Accuracy rvB 95 Feb. June 5-8 3.2 SS.2 S4-3 18.8 7 0 3 8 - 0.9 IVA los Feb. June 6 7 7 9 3 4 5 4 S2.2 55-7 25. 9 3-5 VB 79 Feb. June 8 8 I 3 4 4 S 6 SS-S 55-4 2.2 — 0.1 VA 107 Feb. June 8 8 4 4 5 4 3 9 63.1 58.3 - 7-5 - 4.8 VIB 83 Feb. June 8 9 7 7 S 6 S 9 64 72 2SS 8.0 VIA 83 Feb. June 9 10 0 S 5 7 3 5 59 71 41s 12.0 vnB 63 Feb. June 9 10 7 8 S 8 6 I 60 75 44-6 15.0 VIIA 67 Feb. June 9 11 8 8 6 8 0 4 62 71 40.0 9.0 VIIIB 65 Feb. June II 12 4 0 6 9 9 3 61 78 34-8 17.0 vniA 62 Feb. June II 13 S 7 6 10 3 4 55 76 65.1 21.0 SUBTRACTION IVB 95 Feb. June 5.2 6.7 2.6 4-7 500 70.1 80.0 20.1 IVA 105 Feb. June 5-7 6.7 35 4.2 61.4 62.7 20.0 1.3 VB 79 Feb. June 6.7 7-9 4.6 51 68.9 64. 5 10.9 - 4-4 VA 107 Feb. June 7-3 8.1 4-7 5-3 64.8 65-4 130 0.6 VIB 83 Feb. June 8.0 8.1 55 6.0 69 74 91 5.0 VIA 83 Feb. June 91 91 7-1 7.2 79 79 1-4 0.0 vnB 63 Feb. June 9-7 9 5 7.6 7.6 79 80 0.0 I.O VHA 67 Feb. June 10.4 10. 1 7.6 7.2 73 71 - 5-3 — 2.0 vniB 6S Feb. June IX. 0 10.8 8.8 8.3 80 77 - 5-7 - 30 vmA 62 Feb. June 130 II. 9 10.3 9-7 82 - 5.8 30 A HALF-YEAR'S PROGRESS IN ONE SCHOOL SYSTEM 85 TABLE II — Continued MULTIPLICATION Grade No. OF Pupils Date Attempts Rights Accuracy Percent Gain In Rights IVB . . IVA.. VB... VA... VIB . . VIA.. VIIB. VIM VIIIB \'IIIA IVB.. IVA.. VB... VA... VIB.. VIA.. VIIB. VIIA. VIIIB VIIIA 95 los 79 107 83 83 63 67 6S 62 Feb. June Feb. June Feb. June Feb. June Feb. June Feb. June Feb. June Feb. June Feb. June Feb. June 5-2 6.2 6.3 6.9 6.6 7-1 7.0 7.4 8.0 7.8 8.S 8.7 9.0 9-4 45-2 54-2 46.2 S9-7 61.9 53-6 6s.i 63 -4 64.2 66.7 65.7 63 S 62.5 60. 2 64.7 62 66.7 66.5 61.3 65.7 6S.4 59-2 - 5-1 4.7 2.3 2. 2 - 6.0 - 1.8 3-3 8.1 95 Feb. June los Feb. June 79 Feb. June 107 Feb. June 83 Feb. June 83 Feb. June 63 Feb. June 67 Feb. June 65 Feb. June 62 Feb. June 13 ° 5 1 4 4 I 8 3 8 I S 4 8 3 0 4 6 2 8 5 5 3 2 5 .5 3 7 6 3 4 6 6 0 4 6 6 8 4 3 6 5 5 3 6 7 5 4 7 4 6 0 7 3 6 4 8 8 7 I 8 8 7 5 9 4 8 2 9 2 8 0 10 4 8 9 II I 9 8 38.4 40.9 39-5 62.5 60.9 S6.i 73-5 76.7 63.2 81. 5 80.6 87.7 80.7 85.2 85.6 88.3 260.0 100. o 14-3 243 - 6.5 1.9 6.7 S.6 - 2.4 10. 1 86 THE FIFTEENTH YEARBOOK Grade Score 4B 4A sB s^ 6B 6A 7B 7A SB 8A „__! I 1 1 ~"~" —•"•-— ——— 7 e. - - - e. A d '. ^--. ........ ••"•^•' ■ ■ ^ « * :: : ^i2lt~. :__ ._ _ _ ^'' _ _ _ -_ _ _ ,^ _ __ < *^ „^», .- _^; . — ^ — """K ^ " -^ ^s I -* 2 ^51 t 2 ** ,L IZ. . I ^^^ _^^«-!5»^^__^-^r " T""""ii:^ i IIIIIIIIILMtiitfniill III r^ / ^ii. /,,,± :::::::::::::::::::::.:::::::: i::::!^::::;;^::::: :::::::::::::::: : - s"** ■ir[$" ' " J — 1 1 _ _fli iffiMMMiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiyjj -■^^^ February June Chakt II. — Progress in arithmetical achievement at Bloomington, Indiana. Addition, rights. A HALF-YEAR'S PROGRESS IN ONE SCHOOL SYSTEM 87 — — February June Chaut III. — Progress in arithmetical achievement at Bloomington, Indiana. Subtraction, rights. 88 THE FIFTEENTH YEARBOOK Grade Score 4B 4A 5B sA 6B 6A 7B 7A 8B 8A ^^— February June Chart IV. — Progress in arithmetical achievement at Bloomington, Indiana. Multiplication, rights. A HALF-YEAR'S PROGRESS IN ONE SCHOOL SYSTEM 89 Grade Score 4B 4A 5B 5A 6B 6A 7B 7A 8B 8A -^^— February - June Chart V. — Progress in arithmetical achievement at Bloomington, Indiana. Division, rights. 90 THE FIFTEENTH YEARBOOK are highest in the order of processes as follows: addition, subtraction, multiplication, and division. In Grade VIIIA (save for the drill influ- ence in addition) the order is division, subtraction, addition, and multi- plication— the latter two have about equal rank; (7) that the increase in accuracy from grade to grade is most pronounced in division, ranging from 38 . 4 in the IVB February test to 89 in the VIIIA June test; and is least pronounced in addition (save for the drill influence preceding the June test); the percentage of accuracy as indicated by the February test is 55 in both Grades IVB and VIIIA. The marked gain in accuracy and in rights between the February and the June tests in addition in Grades VI, VII, and \rEII, is accounted for by the fact that a five-minute daily drill in addition was given for a period of ten weeks in the interval between the two tests in these grades.^ An average gain in rights of 42 per cent and in accuracy of 14 per cent in addition, under drill conditions between the February and June tests, in Grades VI, VII, and VIII, indicates that present standards, under conditions of ordinary class work, are no indication of what these standards should be when experimentation has shown the way to a better procediire. The marked gain in results in addition between the February and the June tests and the lack of gains in the other processes in Grades VT, Vn, and VIII indicate that results may be expected at the point where pressure is exerted and that there is no appreciable transfer of training from one process to another. In grades from IVB to VIIIA, the teachers were rated as to their general efficiency by four judges and the averages of these ratings were correlated with the amount each teacher's class gains in arithmetic were above or below the median class gain, as measured by the February and June tests. In Grades IV and V, where the teachers have charge of all subjects, the Pearson coefficient was about 4-0.30, but in the depart- mental grades, where teachers of arithmetic are employed because of their proficiency in teaching this subject, the coefficient of correlation was above -f-o . 90. These results are offered as data from one school system only and are not to be considered as determined general standards of achievement and growth in arithmetical abilities. ' Acknowledgments are due to Miss Mary Kerr, principal of the departmental school, who planned and carried out the drill work in addition and assisted in the tabulation of results. CHAPTER VIII COURTIS TESTS IN ARITHMETIC: VALUE TO SUPERINTENDENTS AND TEACHERS S. A. COURTIS Supervisor of Educational Research, Detroit, Michigan From August i, 1914, to August i, 1915, between four and five hun- dred thousand tests (455,007) of the various Courtis standard research tests were sent from Detroit to school men in 42 different states. This material was mainly Series B Arithmetic Tests, and the growth from the use of the tests by a single school in 1909 may be taken as an index of the growth throughout the United States of the interest in the movement for measurement. It should be evident at once that if this great quantity of material is being used so as to result in benefit to the schools tested, then measurement must be already exerting throughout the country a very widespread influence on the teaching of arithmetic. On the other hand, this extensive use of testing material may represent merely natural curiosity and an experimental trial by wide-awake school men of a new and much-discussed type of examination. Fortunately, measurement itself is not on trial. The movement for measurement is merely an application of scientific methods to the study of educational problems. Both as a general method for the dis- covery of natural law and as a method of proved worth in education, the method of science rests on so sure a foundation that for a school man to declare that in his hands measurement has been a failure is to confess his own lack of training or his own incompetency. As for the Courtis tests, they were designed by the writer for a specific purpose, viz., to measure the effects of his own teaching and of methods invented to improve its efl&ciency. This purpose has been successfully accomplished. For the writer's own purposes the tests yield results which are satis- factory and which have fully justified the time and effort given to the testing work. Further, the returns received from other schools to which the tests were sold on the co-operative basis have yielded informa- tion which has proved of the highest professional interest and value to 91 92 THE FIFTEENTH YEARBOOK him. Moreover, the tests have given satisfactory service in the hands of so many professors and students of education that their value as tools for educational research is well established. But whether or not either the method or the tests are of value to superintendents and teachers generally is quite a different question. Accordingly, at the request of the chairman of the Committee on Standards and Tests of Efficiency, an investigation of this question was undertaken. The following letter was sent to 200 superintendents in 30 states, all of whom had recently purchased copies of the Series B Arithmetic Tests: Dear Sir: I am in receipt of a request from the Committee on Standards of the N.E.A. asking for the conclusions of superintendents as to whether or not the use of the Courtis Standard Tests in Arithmetic has been of any value to them or to their teachers from the standpoint of school administration or teaching. Any statement that you may be willing to send me will be forwarded to the committee. Thanking you for such assistance as you may be willing to give, I am, Yours very truly, S. A. Courtis At this writing replies have been received from S7 superintendents in 30 states and they are still coming in. The states represented are Alabama, California, Colorado, Connecticut, Illinois, Indiana, Iowa, Kansas, Michigan, Minnesota, Missouri, Montana, Nebraska, New Hampshire, New Jersey, New York, North Carolina, Ohio, Oklahoma, Oregon, Pennsylvania, Rhode Island, Tennessee, Texas, Utah, Virginia, Washington, West Virginia, Wisconsin, and Wyoming. The general tone of these letters is remarkable for its enthusiastic commendation of the value of measurement to superintendents, teachers, and pupils. Such expressions as "Delighted with results," "Should not like to do without them," "Do more than anything else we have ever tried," or their equivalents, occur in the majority of the letters. Seven of the eighty-seven stated that their results have not been tabu- lated because of lack of time. Only two express dissatisfaction. The following letter represents the extreme of unfavorable comment: Replying to yours of the 29th ultimo, I regret to say that I have not dis- covered any material benefit from the Coiurtis tests as we applied them last year. COURTIS TESTS IN ARITHMETIC 93 If one had a large amount of statistical assistance they might be worth while, but as a superintendent with a limited force I question their value. Yours sincerely, [The name is omitted for obvious reasons.] On the other hand, many of the other letters show a careful study of the problem and a thorough analysis and formulation of the benefits. The following letter is an illustration: Burlington, Iowa, October 4, 19x5 Mr. S. A. Courtis 82 Eliot Street Detroit, Mich. Dear Sir: The Courtis Standard Tests in Arithmetic have been of great value to me in indicating: I. The school or schools in which there is a regular increase in abUity of the pupils in the fundamentals of arithmetic. Such knowledge has enabled me to suggest to the principals in those buildings where there is no such regvdar increase in ability of the pupUs methods of bettering the product of their teaching. II. The room or grade in which there has been no increase in ability during the semester. This indicates in some measure the work which the teacher is doing. It affords the superintendent an excellent basis for discussing in detail with the teacher the faults or good points in her work. III. The school or grades in which there is a tendency to emphasize the work of the fundamentals beyond what is reasonable. Teachers like to do what they can do well. Their interest in those subjects which they like some- times carries them too far. From the statements made to me by teachers and principals I am confident that the tests have not produced as good results as they should, but this is not the fault of the tests. As the teachers come to imderstand the purpose of the tests, the value of this work becomes more and more apparent to all. I do not think that we care to give up the tests under any circumstances. Yours very truly, W. L. Hanson, Superintendent It is to be regretted that space prevents quotations from many others. Replies were tabulated, as follows: Replies Total replies received 87 Tabulations not completed 7 Unfavorable 2 Favorable 78 94 THE FIFTEENTH YEARBOOK Benefits Mentioned General answers only 17 Reveal needs of individuals 23 Comparison with other cities .... 14 Stimulate teachers 23 Comparison from grade to grade. . 7 Stimulate pupils 17 Reveal weak points in school work 1 1 Furnish standards 24 Reveal weak teachers 13 Furnish incentive or motive 16 The results do not lend themselves to statistical tabulation, owing to the variation in the forms used in expressing the same ideas. Other important uses appearing in these letters are "Measurement of the efficiency of various methods," "Measurement of the progress of indi- vidual children," "Value in grading," "Awakening of spirit of investiga- tion among teachers," "Satisfaction of parents." The conclusions to be drawn from the foregoing data are that for the most part the superintendents who applied for standard tests are making legitimate use of them for purposes of supervision; that these men value the comparison from city to city — made possible by imiform tests and conditions — as a check upon the main character of the work done in their own systems; that the use of standard tests results in the setting up of objective standards which affect the work of teachers and pupils favorably, both by making clear the goal to be attained and by furnishing motives for individual effort; that the tests are of great value in the determination of the needs of individual children and in the adjustment of school work in arithmetic to such needs; that the tests have some value, the amount of which is yet to be determined, in the judging of the efficiency of teachers and in determining the grades of children; that few superintendents are making use of the tests in a scientific study of the comparative efficiency of different methods of teaching. The rapid increase in the number of tests that have been used each year is, therefore, probably due more to the value of the results secured than to mere curiosity in a passing fad. The use of standard tests for purposes of supervisory control under such conditions is sure eventually to have profoimd influence upon the teaching of arithmetic. The writer is glad to have this chance to express his appreciation of the co-operation of the many school workers who have made this exten- sive experimental use of the research material possible. The financial burden and office labor of carrying on, without profit, a co-operative COURTIS TESTS IN ARITHMETIC 95 venture totaling some $5,000 or S6,ooo a year has not been light; neither has the labor of tabulation of returns. It is therefore gratifying to find that this work has been of real value to many superintendents, and the result of the investigations made for the committee will be to stimulate further efiforts to secure standards of teaching efl&ciency and to extend the range of the testing material. VALIDITY OF STAISTDARD SCORES The first tests in arithmetic were issued in 191 1 and distributed widely in an attempt to secure standards for use in the writer's own classes. The first tabulation of the returns obtained was in June, 1911, and every year since that date additional tabulations have been made. Series B tests were issued at the beginning of the year 19 13-14, and standards based upon the first tabulations were issued in February, 1914. At this time, however, both the tests and the method had reached a stage of development which made possible effective work, so that no change in standards has been necessary since that time, although tabula- tions of larger and larger numbers of scores have been repeatedly made. The standard scores set for Series B are as shown in Table I. TABLE I Standard Qune) Scores. Series B Tests Grade Test I Addition Test 2 Subtraction Test 3 Multiplica- tion Test 4 Division m IV 3 5 7 9 II 12 4 6 8 10 II 12 3 5 7 9 10 II 2 4 6 V VI VII VIII 8 10 II Standard accuracy = ioo per cent. The scores given in Table I represent approximately the median speed of work for the different grades and are based upon returns that are nearly nation-wide in scope. The range of variation in schools in different cities and states is approximately four examples above and below the median; i.e., in some school systems the median eighth-grade scores will rise as high as 16 examples in addition and others go as low 96 THE FIFTEENTH YEARBOOK as 8 examples. Not more than five eighth-grade classes per hundred will exceed these limits, except as very peculiar and special conditions prevail. On the other hand, the range of speed of work in individuals varies from a score of but two or three examples to scores of twenty-four examples, the limit of the test. The conditions from city to city do not show greater variation in achievement than are to be found in any one city, such as Boston or Detroit, where there is a large number of classes of the same grade. If returns from small cities or country schools only were tabulated, the median scores for any given grade would probably tend to be somewhat lower. The large-city school apparently emphasizes the drill work. The problem of setting of an adequate standard is, therefore, a difficult one. Any standard adopted must take into consideration the effect of a number of different factors. All things considered, it has seemed best to take as a standard of speed the median speed derived from tabulations of all t3^es of systems. There should certainly be no attempt to press training in addition, for instance, to very high levels of ability at the expense of more important work, and very few school men are willing to neglect in any way training in such fundamental abilities as the four operations. Median speed, determined from a wide range of conditions, probably represents the optimal speed at which children can work. A trial in the classroom of such speeds as standards has yielded satisfactory results. The question of standards of accuracy, however, is a much more difficult one to settle, because less information is available and there is more room for a play of personal opinion. The writer has as yet reached no conclusion in the matter, but is endeavoring to determine the degree of accuracy which it is practical to attain under classroom conditions. For this purpose it is necessary to set before teachers as a goal to be reached the highest ideals possible — i.e., loo per cent accuracy — then to determine in terms of the percentage of the class reaching this goal the degree of success which it is possible to attain. For instance, the average percentage of children of the eighth grade who show median speed with loo per cent accuracy in first draft work is between 5 and 10 per cent. Experiment proves, however, that it is easily possible to raise the group showing perfect accuracy to 20 or 30 per cent of the class membership and markedly to increase the number of children working with accuracies of 90 and 80 per cent. There is even reason to COURTIS TESTS IN ARITHMETIC 97 expect that with proper methods of training and by emplojdng standards throughout the whole school system, and without change in the time given to arithmetic, it will be possible eventually to secure perfect accuracy in from 60 to 75 per cent of the children. For the classes under his immediate control the writer prefers to keep the standards of median speed and 100 per cent accuracy as the goals to be attained. He recog- nizes clearly, however, that at present this choice of standards must rest upon personal convictions only, and school men should feel free to change these standards to suit their own opinions. There are, however, certain facts, other than the achievements of the pupils themselves, which ought to be considered in the determining of standards. One of these is the social value of the abihties developed by school work. The writer has attempted to answer this question by the measurement of as many adults as possible. The first attempt along this line was made in connection with a survey of the New York schools for the Committee on School Inquiry. Through the kindness and co-operation of Mr. W. D. Ernest, chief of Cadet Staff and member of the John Wanamaker New York Commercial Institute, the consent of Mr. Lynn, general manager of the John Wana- maker department store. New York City, was secured to the testing of 50 employees of the company. This group was tested precisely as if it had been a class of children in school. It met in one of the company's schoolrooms and was tested by one of the force of trained examiners used in the New York survey. Exactly the same tests and time allowances were used as for the children and the same procedure in conducting the examination and in scoring and tabulating the papers was followed throughout. Forty-one complete records were obtained. The subjects represented six different types of positions in the store and in numbers were as follows: Auditing department 5 Salesmen 7 Billing clerks 5 Typists 3 Cashiers 8 Clerks 13 Total 41 Two of the clerks and six of the sales people were men. The average age of the group was approximately nineteen years, ranging from fifteen to thirty years. The average term of service for the company, except for the group from the auditing department, was a Httle more than two years, ranging from two months to five years. The girls from the auditing 98 THE FIFTEENTH YEARBOOK department are the product of the store's own training and the term of service for them ranged from eight to fourteen years. The wage paid any member of the group is determined by position and term of service, not by position alone. The amount ranged from five to fifteen dollars per week. Of thirty-six who reported the last grade attended in pubhc school, seven SJ 01 o 1^ S 6 D 3 ^ , , ts -M ra s a n < W Fundamentals- -Attempts Chart I. — Line marked S indicates eighth-grade standard, deviations above and below standard. Rights Scale at left shows gave high school, thirteen had completed the elementary grades, and sixteen were in either the seventh or the eighth grade of the grammar school when they left. It was not possible to attempt more than a general study of the work of the different groups. The cashiers do little more than make change; the clerks and salesmen have a little computation work in the handling of sales slips, store records, etc.; the members of the auditing department have a larger amount of abstract work and it is routine in COURTIS TESTS IN ARITHMETIC 99 character. The auditing department and, to a lesser extent, the billing clerks are thus the only positions in which arithmetical ability would have more than a slight influence in determining the fitness of the applicant. The tests used were Series A, and in Table II the results are given for Test 7 only, a test in the four operations with whole numbers. The scores given in this test do not differ markedly from those of Series B; the eighth-grade standard score is about 2^ examples higher. The results of the tests are given in Table II and shown in the graph (Chart I; see p. 98). TABLE II Part B. Average Scores Position Salesmen. . . . Cashiers .... Typists Clerks BUI clerks . . . Auditors. . . . Entire group . Rights * Computed scores. Entire tests finished in less than time allowed It will be noted that the scores of salesmen and cashiers fall below the eighth-grade standards (14.4 examples attempted, 10 examples right); those of the typists and clerks are almost exactly at standard; and those of the billing clerks and of the auditing department run con- siderably above standard. Two members from the auditing department had very high scores; the best one finished the test in so short a time that had enough material been furnished to keep her busy during the whole time allowed, her score would have been 38 attempted and 34 right. Before commenting on these results, however, similar results from other sources will be presented. Through the co-operation of Miss Ade- laide Baylor, clerk of the Board of Education for the state of Indiana, and the kindness of Mr. Jesse Moore, president of the Columbia School Supply Company, tests were given to a group of 66 factory laborers. These workmen were of three classes: a group of 20, mostly colored men, represented the cheapest labor employed in the factory, average lOO TEE FIFTEENTH YEARBOOK wage lo to 12 cents per hour; a second group of 26 men represented the median wage in the factory, average 17.6 cents an hour; a third type represented the best labor in the factory outside of the ofi&ce, average wage 21 cents an hour. Tests were also given to a group of 13 sales- women, ranging in age from nineteen to thirty-six. The average score made by these four groups of employees is shown in Table III. Through the kindness of personal friends and the co-operation of Mr. Boyd Fisher, secretary of the Executive Club of the Detroit Board of Commerce, additional records have been secured from various t}^es of adults. It is difficult, however, to obtain complete data in such cases, as information in regard to either age, salary, or occupation is likely to be missing. However, in Table III will be found the records of a group of low-wage girls ranging in age from eighteen to twenty-five, with average pay of $400 per year; 3 stenographers, age eighteen; 5 adult women, ranging in age from thirty-seven to forty-six, who give their occupation as housewife; a group of high-wage women ranging in age from twenty-one to forty and in salary from $700 to $1,200 per year; 14 boys and men representing machinists, steam-fitters, bookkeepers, railway, foremen, railway clerks, trimmers, and salesmen, ranging in age from nineteen to forty-four and in salary from $350 to $900 a year; a group of 7 high-priced men of independent means, ranging from thirty- seven to fifty-nine years of age; a group of 44 Iowa superintendents, ranging in salary from $800 to $4,500, and a similar group of Michigan superintendents, all about thirty-five years of age, and ranging in salary from $600 to $3,600; a group of 17 office employees of an automobile company in Detroit, ranging in age from eighteen to thirty-two, and in salary from $700 to $1,500 per year; a group of 28 employees of the City Gas Company, ranging in salary from $300 to $5,000 per year; a group of approximately 80 teachers, mostly women, attending the sum- mer school of the George Peabody College for Teachers, ranging in age from twenty to forty-five. In Table IV, the individual scores of one of the groups are given in full. It is evident from these tables that there is an apparent correlation between the earning capacity of adults and their scores, but whether this is a causal relation or not is another question. The fact that a man attains a high position in society is more likely to be due to the superior quality of his general abilities than to his ability in arithmetic alone. An able individual will profit more by school training than one less gifted, COURTIS TESTS IN ARITHMETIC lo:; >< PQ 1— 1 1— 1 H t— 1 ^ § pq < H H CO sjqSiH sjdraa^jv XoBinDDy siqSrg siduia^^v }U33 jaj sjqSi^ sidinanv •* M vO 00 o O H W 00 cs t^ Ov O O too ooO vO lOOO O t^oo woe* \0 r-» M -^oo Tj-O 00 O 0^ rC O 00 00 O oo (N POO Ov\0 lO r^ -^ 0> PO O POO P< On t^oo O O t^ o M -^O O t^ Ov t^ lo O O O 1-1 »^ O O N 0\0 o o t~ ■>* p< lo c* « H 00 tS 00 O H 0 rj- r^ PO cs PI On M '*• s?q3rg s?dni3;iy 0> O oo On O 00 ■* Tl-oO On On vo vo O M PO On M M to PO OnO O ■* O •>* P< O 00 H N M r^ cs 00 t^ P) O N 'i- •*00 O O STvnoiAiaNi o o o o o o o o to On too itii O to O O r^ pooo O I I 00 00 r^ w w PO to O O PJ Tl- Tf Tj- PO fO •* I I I I 1 O l-l OnOO On S-2 ^^ o'^ On pO PO O O NO 00 t^ o H O PO O PO to to MO o to to to o o od 00 o S " 6 cj >. tn 2-C S a § W).^ J^ fe f O > Ui C3 C3 l^ls o ^ cn-§i^r-i 5 o « "* J. ? O IJ -C ?3 u u : c tn 1) c y v» [Checking accoimt 251 . 12 Total $2,974.07 MEMBERSHIP Nimiber of active members (including one honorary) December 23, 191S IQS Nimiber of associate members December 23, 1915 242 Total membership 437 Guy M. Whipple, Secretary-Treasurer LIST OF ACTIVE AND HONORARY MEMBERS OF THE NATIONAL SOCIETY FOR THE STUDY OF EDUCATION (Corrected to January 15, 1916) HONORARY MEMBERS Dewey, John, Columbia University, New York, N.Y. ACTIVE MEMBERS Allen, Fiske, State Normal School, Charleston, 111. Arbaugh, W. B., Ypsilanti, Mich. Ashley, L. S., Sibley, lU. Axline, Howard E., West Technical High School, Cleveland, Ohio. Bachman, R. H., Tarboro, N.C. Bagley, William C, University of Illinois, Urbana, 111. Baldwin, Bird T., Swarthmore College, Swarthmore, Pa. BaUou, Frank W., care of School Committee, Mason St., Boston, Mass. Barnes, Harold, Girard College, Philadelphia, Pa. Becker, Ernest J., Eastern High School, Baltimore, Md. Bell, J. Carleton, University of Texas, Austin, Tex. Benedict, Ezra W., Principal of High School, Walden, Orange Co., N.Y. Blaine, Mrs. Anita McCormick, loi East Erie St., Chicago, lU. Bland, W. P., Superintendent of Schools, Globe, Ariz. Bolton, Frederick E., University of Washington, Seattle, Wash. Boyer, Charles, Superintendent of Schools, Atlantic City, N.J. Bradford, Mrs. Mary D., Superintendent of Schools, Kenosha, Wis. Breckenridge, Elizabeth, 962 Fourth St., Louisville, Ky. Briggs, Thomas H., Columbia University, Teachers College, New York, N.Y. Brooks, Stratton D., State University, Norman, Okla. Brown, J. C, University of Illinois, Urbana, 111. Brown, J. Stanley, Superintendent of Township High School, Joliet, 111. Brown, John F., 559 W. 156 St., New York, N.Y. Brumbaugh, Martin G., Harrisburg, Pa. Bryan, W. J. S., Assistant Superintendent of Schools, St. Louis, Mo. Buchner, Edward F., Johns Hopkins University, Baltimore, Md. Bumham, Ernest, State Normal School, Kalamazoo, Mich. Bush, I. B., Superintendent of Schools, Erie, Pa. Butterworth, Julian E., University of Wyoming, Laramie, Wyo. 167 l68 THE FIFTEENTH YEARBOOK Cammack, 1. 1., Superintendent of Schools, Kansas City, Mo. Chadsey, Charles E., Superintendent of Schools, Detroit, Mich. Chandler, J. A. C, Superintendent of Schools, Richmond, Va. Chapman, Ira T., Superintendent of Schools, Norwalk, Conn. Charters, W. W., State University, Columbia, Mo. ChUds, Hubert G., State University, Bloomington, Ind. Clark, W. A., 312 E. Jefferson St., Kirksville, Mo. Claxton, P. P., Bureau of Education, Washington, D.C. Clement, J. A., Northwestern University, Evanston, 111. Coffman, Lotus D., University of Minnesota, Minneapolis, Minn. Condon, Randall J., Superintendent of Schools, Cincinnati, Ohio. Conradi, Edward, Florida State College for Women, Tallahassee, Fla. Cook, Albert S., County Superintendent of Schools, Sta. A., Towson, Md. Cook, John W., President, Northern Illinois State Normal School, DeKalb, 111. Cooke, Flora J., Francis W. Parker School, 330 Webster Ave., Chicago, 111. Courtis, S. A., 82 Eliot St., Detroit, Mich. Cox, Philip W. L., Superintendent of Schools, Solvay, N.Y. Cramer, W. F., Superintendent of Schools, Atlantic, la. Cubberley, Ellwood P., Leland Stanford Junior University, Stanford University, Cal. Davis, B. M., Miami University, Oxford, Ohio. Davis, Solon P., District Superintendent, Henry Barnard School, Hartford, Conn. Deahl, Jasper N., University of West Virginia, Morgantown, W.Va. Dearmont, Washington S., President, State Normal School, Cape Girardeau, Mo. Dick, George S., President, State Normal School, Kearney, Neb. Doelle, John A., Superintendent of Schools, Houghton, Mich. Doyle, Mary E., 1009 60th St., Chicago, 111. Earhart, Lida B., 430 W. ii8th St., New York, N.Y. Eby, Frederick, State University, Austin, Tex. Elliott, C. H., 527 W. i2ist St., New York, N.Y. Elliott, Edward C, Chancellor, State University, Helena, Mont. Ellis, A. Caswell, University of Texas, Austin, Tex. Elson, William H., Superintendent of Schools, Cleveland, Ohio. Ernst, L. R., Cote Brilliante School, St. Louis, Mo. Farrington, Frederick E., Teachers College, Columbia University, New York, N.Y. Feeney, T. L., Miami University, Oxford, Ohio. Felmley, David, President, Illinois State Normal University, Normal, 111. Fleshman, A. C, 1937 South Grand Ave., Los Angeles, Cal. LIST OF ACTIVE AND HONORARY MEMBERS 169 Forbes, George M., 235 Dartmouth St., Rochester, N.Y. Foster, H. H., University of Arizona, Tucson, Ariz. Frederick, J. M. H., Superintendent of Schools, Cleveland, Ohio. Frost, J. M., Superintendent of Schools, Muskegon, Mich. Fuerst, Sidney M., 176 Amity St., Flushing, Long Island, N.Y. Gosling, T. W., Hughes High School, Cincinnati, Ohio. Greeson, William A., Superintendent of Schools, Grand Rapids, Mich. Grifl&n, Margery M., 4045 McPherson Ave., St. Louis, Mo. Gwinn, J. M., Superintendent of Schools, New Orleans, La. Hall, John W., University of Cincinnati, Cincinnati, Ohio. Halleck, Reuben Post, Principal, Boys High School, Louisville, Ky. Hamilton, Cora M., State Normal School, Macomb, lU. Harris, Ada Van Stone, 6216 Howe St., E.E., Pittsburgh, Pa. Hatch, W. H., Superintendent of Schools, Oak Park, 111. Heatwole, Cornelius J., State Normal School, Harrisonburg, Va. Heckert, J. W., Miami University, Oxford, Ohio. Henderson, Hermon C, State Normal School, Milwaukee, Wis. Herron, Helen, 1933 Elysian Fields Ave., New Orleans, La. Hill, Patty Smith, Teachers College, New York, N.Y. Hitchcock, Clara M., Lake Erie College, Painesville, Ohio. Horn, Paul Whitfield, Superintendent of Schools, Houston, Tex. Jefifers, Fred A., Supermtendent of Schools, Painsdale, Mich. Jenks, Jeremiah W., New York University, Washington Square, New York N.Y. Johnson, Pliny, Woodward High Schools Cincinnati, Ohio. Johnston, Charles Hughes, University of Illinois, Urbana, 111. Jones, Arthur J., State University, Orono, Me. Jones, Lewis H,, President, State Normal College, Ypsilanti, Mich. Judd, Charles H., University of Chicago, Chicago, 111. Kemp, W. W., University of Montana, Missoula, Mont. Kent, Henry L., Agricultural College, Manhattan, Kan. KimbaU, J. F., Superintendent of Schools, Temple, Tex. King, John P., Superintendent of Public Schools, Union City, Ind. Kirk, John R., President, State Normal School, Kirksville, Mo. Kirk, W. H., Superintendent of Schools, East Cleveland, Ohio. Latham, R. H., Winston-Salem, N.C. Lattimore, J. C, Superintendent of Schools, Waco, Tex. Laurence, Isabel, State Normal School, St. Cloud, Minn. Lawson, W. C, Superintendent of Schools, Bryan, Tex. Lewis, Homer P., Superintendent of Schools, Worcester, Mass. Logan, Anna E., Miami University, Oxford, Ohio. I70 THE FIFTEENTH YEARBOOK Lord, L. C, State Normal School, Charleston, 111, Lowiy, Charles D., 1643 Kenilworth Ave., Chicago, 111. Lucas, Hardin, State Normal School, Valley City, N.Dak. Luckey, G. W. A., State University, Lincoln, Neb. Lnkens, Herman T., Francis W. Parker School, 330 Webster Ave., Chicago, 111. Lull, H. G., State University, Seattle, Wash. McAlUster, Cloyd, N., Berea, Ky. McDaniel, C. JNI., Superintendent of Schools, Hammond, Ind. McKenney, Charles, State Normal School, Ypsilanti, Mich. Mackey, E., Superintendent of Schools, Trenton, N.J. McMurry, Charles A., State Normal School, DeKalb, 111. McMurry, Frank M., Teachers College, New York, N.Y. Manny, Frank A., Teachers Training School, Baltimore, Md. Maphis, Charles G., University of Virginia, University, Va. Marrs, S. M. N., Superintendent of Schools, Terrell, Tex. Marsh, J. F., Assistant State Superintendent of Schools, Charleston, W.Va. Marsh, M. E., Berea College, Berea, Ky. Maxwell, William H., Superintendent of Schools, New York, N.Y. Melcher, George, 9th and Locust Sts., Kansas City, Mo. Miller, Irving E., University of Rochester, Rochester, N.Y. Miller, James C, Department of Education, Edmonton, Alberta, Can. Mills, Harriette M., New York University, Washington Sq., New York, N.Y. Minnich, H. C, State Normal College, Oxford, Ohio. Monroe, Edwin S., Superintendent City Schools, Muskogee, Okla. Monroe, Walter S., Kansas State Normal School, Emporia, Kan. Morgan, O. S., Alfred University, Alfred, N.Y. Morrison, H. C, State Superintendent of Schools, Concord, N.H. Newton, George A., Trinity University, Waxahachie, Tex. Nichols, C. A., Southwestern University, Georgetown, Tex. O'Donnell, James A., 95 Boervmi St., School 43, Brooklyn, N.Y. Olin, A. S., State University, Lawrence, Kan. O'Shea, M. V., University of Wisconsin, Madison, Wis. Pahner, G. Lloyd, Superintendent of Schools, Frederick, Md. Parker, S. Chester, University of Chicago, Chicago, 111. Peterson, O. E., Sycamore, 111. Phelan, W. W., 629 Boyd St., Norman, Okla. Pollack, Rosalie, University of Nevada, Reno, Nev. Pusey, E. D., Superintendent of Schools, Durham, N.C, Putnam, Helen C, Rhode Island Ave., Providence, R.I. Rail, E. E., State University, KnoxviUe, Tenn. LIST OF ACTIVE AND HONORARY MEMBERS 1 71 Rapeer, Louis W., Pennsylvania State College, State College, Pa. Reigart, J. F., 31 Euclid Ave., Yonkers, N.Y. Rosier, Joseph, Superintendent of Schools, Fairmont, W.Va. Roy, Victor L., President, State Normal School, Natchitoches, La. Ruediger, W. C, George Washington University, Washington, D.C. Russell, James E., Dean of Teachers College, New York City, N.Y. R3aiearson, Edward, Principal, Fifth Avenue High School, Pittsbiurgh, Pa. Sachs, Julius, Columbia University, New York City, N.Y. Scott, Z. E., City Superintendent, Asbury Park, N.J. Slauson, Herbert M., Superintendent of Schools, Ann Arbor, Mich. SmUey, William H., Superintendent of Schools, Denver, Colo. Smith, H. L., Superintendent of Schools, Bloomington, Ind. Smoot, Lucy J., 401 1 Baltimore St., Kansas City, Mo. Snedden, David S., 302 Ford Bldg., Boston, Mass. Snyder, Z. X., President, State Normal School, Greeley, Colo. Starch, Daniel, State University, Madison, Wis. Stark, William E., Supervising Principal of Schools, Hackensack, N.J. Stillwell, WUliam E., University School, Cincinnati, Ohio. Stockard, L. V., Austin High School, Austin, Tex. Stone, Cliff W., Iowa State Teachers College, Cedar Falls, la. Stowe, A. Monroe, Toledo University, Toledo, Ohio. Strayer, George D., Teachers College, Columbia University, New York, N.Y. Strong, B. Norman, Arsenal School, Hartford, Conn. Sutton, W. S., University of Texas, Austin, Tex. Suzzallo, Henry, State University, Seattle, Wash. Taft, Leanora E., Woodstock, Vt. Taylor, Joseph S., 2275 Loring Place, The Bronx, New York, N.Y. Thompson, Frank E., University of Colorado, Boulder, Colo. Thomdike, Edward L., Teachers College, Columbia University, New York, N.Y. Thurber, Charles H., Editor, Ginn & Co., 29 Beacon St., Boston, Mass. Updegraff, Harlan, University of Pennsylvania, Philadelphia, Pa. Vanderwalker, Nina C, State Normal School, Milwaukee, Wis. Van Sickle, James H., Superintendent of Schools, Springfield, Mass. Vincent, H. D., Principal, Public School 3, Troy, N.Y. Waldo, Dwight B., Western State Normal School, Kalamazoo, Mich. Walker, Elmer W., Superintendent, State School for Deaf, Delavan, Wis. Weber, S. E., Superintendent of Schools, Scranton, Pa. Weglein, David E., Western High School, Baltimore, Md. West, Henry S., University of Cincinnati, Cincinnati, Ohio. Whipple, G. M., University of Illinois, Urbana, 111. 172 THE FIFTEENTH YEARBOOK Wiles, Ernest P., Principal, Junior and Senior High School, Evansville, Ind. Williams, Henry G., Ohio University, Athens, Ohio. Wilson, G. M., Iowa State CoUege, Ames, Iowa. Wilson, H. B., Superintendent of Schools, Topeka, Kan. Wilson, Mrs. L. L. W., Philadelphia Normal School, Philadelphia, Pa. Wright, Robert H., Teachers Training School, Greenville, N.C. Zinninger, George Edward, 6i Dewey Ave., Youngstown, Ohio. ggURN EDUCATION-PSYCHOLOGY LIBRARY »ti^-#' 2600 Tolman Hall 642-4209 LOAN PERIOD 1 1 MONTH ALL BOOKS MAY BE RECALLED AFTER 7 DAYS 2- hour books must be renewed in person Return to desk from which borrowed DUE AS STAMPED BELOW JUN Q 6 2m FORM NO. DDIO UNIVERSITY OF CALIFORNIA, BERKELEY I Om, 11/78 BERKELEY, CA 94720 LD 21-50»n-6,'59 (A2845sl0)476 ®i University of Califoornia Berkeley U.C. BERKELEY LIBRARIES CDDSbBMb7Q