Southern Branch 
 of the 
 
 University of California 
 
 Los Angeles 
 
 Form L 1 
 
 LB 
 5063 
 036 
 
 COD.l
 
 JA!> 
 
 This book is DUE on the last date stamped below 
 
 
 JUL 1 9 19 
 
 MAY 5 
 
 2 1 13 <g 
 
 SEP 2& 1942 
 
 JAN 2 9 1946 
 
 MAY Uw 
 
 JAN2o^4S 
 ' AUG 1 2 ^50 
 JUL 3 1953 
 
 AUG 21 ; ~ n 
 
 Form L-9-2»M2,'23 
 
 /
 
 ', 9-" 
 
 r 
 
 THE 
 SCIENTIFIC MEASUREMENT 
 OF CLASSROOM PRODUCTS 
 
 
 J. CROSBY CHAPMAN 
 
 B.A. (Cantab), D.Sc. (London), Pn.D. (Columbia) 
 
 ASSOCIATE PROFESSOR OF EXPERIMENTAL EDUCATION 
 WESTERN RESERVE UNIVERSITY 
 
 GRACE PREYER RUSH 
 
 M.A. (Western P.esebve University) 
 
 INSTRUCTOR IN PHILOSOPHY, WESTERN RESERVE 
 UNIVERSITY 
 
 SILVER, BURDETT & COMPANY 
 BOSTON NEW YORK CHICAGO
 
 Copyright, 1917 
 By SILVER, BURDETT & COMPANY
 
 c 
 
 PREFACE 
 
 The aim of this book is stated in detail in the intro- 
 duction. In short, it is an attempt to present, free from 
 its usual accompaniment of statistical methods, the new 
 idea of educational measurement by means of objective 
 scales. This book will not satisfy the statistically trained 
 reader, for to such its methods will appear clumsy and 
 inelegant. However, not every member of the teaching 
 profession can be expected to have statistical training ; yet 
 it is essential that every one in school work have the quan- 
 titative point of view. If this small book succeeds in in- 
 troducing the reader to the new movement of which the 
 objective scales are the product, it will in a large measure 
 make up for its obvious shortcomings. 
 
 It is our pleasure to record our obligations to the fol- 
 lowing for their courtesy in permitting us to quote their 
 original papers and reproduce their scales : 
 
 Dr. L. P. Ayres (Handwriting and Spelling Scales). 
 
 Dr. F. W. Ballou (English Composition Scale). 
 
 Dr. B. R. Buckingham (Spelling Scales). 
 
 Mr. S. A. Courtis (Arithmetic, Writing, and Reading 
 Scales). 
 
 Mr. W. S. Gray (Reading Scale). 
 
 Dr. M. B. Hillegas (English Composition Scale). 
 
 Dr. Daniel Starch (Reading and Spelling Scales). 
 
 Dr. E. L. Thorndike (Handwriting, Reading, and 
 Drawing Scales). 
 
 Dr. M. R. Trabue (Language Scale). 
 
 Dr. C. Woody (Arithmetic Scale). 
 
 iii
 
 iv Preface 
 
 The full references to these original papers are given 
 in the Appendix. We are also under obligation to Dr. 
 H. Austin Aikins and Miss Myra E. Hills for reading 
 parts of the manuscript. 
 
 One of the authors wishes to express his great indebted- 
 ness to Dr. E. L. Thorndike, from whose writings the 
 ideas embodied in this book largely originated. 
 
 
 J . KJ. O. 
 
 
 G. P. R 
 
 Western Reserve University, 
 
 
 1917. 

 
 TABLE OF CONTENTS 
 
 PAGE 
 
 Introduction. The Aim of the Book . . . . vii 
 
 CHAPTER 
 
 I. Objective versus Subjective Scales op Measure- 
 ment in the Classroom 1 
 
 II. Scales for the Measurement of Ability in 
 
 Arithmetic 9 
 
 III. Scales for the Measurement of Ability in Hand- 
 
 writing 31 
 
 IV. Scales for the Measurement of Ability in 
 
 Reading 63 
 
 V. Scales for the Measurement of Ability in 
 
 Spelling 107 
 
 VI. Scales for the Measurement of Ability in Eng- 
 lish Composition 131 
 
 VII. Scale for the Measurement of Language Ability 157 
 
 VIII. Scale for the Measurement of Ability in 
 
 Drawing 164 
 
 IX. The Application of the Scales in the Schools . 173 
 
 X. Dangers Incidental to the Use of These Scales 182 
 
 APPENDIX A. Sources from Which the Full Account 
 
 of the Scales Can Be Obtained . 189 
 
 APPENDIX B. Bibliography (limited) . . . .191
 
 "Are you content now?" said the Caterpillar. 
 
 "Well, I should like to be a little larger, if you wouldn't 
 mind," said Alice. "Three inches is such a wretched 
 height to be." 
 
 "It is a very good height indeed ! " said the Caterpillar 
 angrily, rearing upright as it spoke (it was exactly three 
 inches high). 
 
 — Lewis Carroll. Alice in Wonderland.
 
 INTRODUCTION 
 
 THE AIM OF THE BOOK 
 
 It may safely be said that the greatest contribution 
 which has been made to education in the last ten years 
 is the application of scientific measurement to school 
 products. The new educational method which has re- 
 sulted shows a clear recognition of the scientific spirit. 
 That a new method was needed was generally agreed, 
 and it has been accepted by those who have experimented 
 with it. However, there is always danger when those 
 engaged in the practice of a subject are for any reason 
 unacquainted with the latest advances. This is clearly 
 demonstrated in a field such as medicine, where it often 
 takes years for a principle which has been accepted by the 
 leaders in the profession to be put into practical applica- 
 tion by the general practitioner. Particularly is this true 
 in the field of education, for any desired advance depends 
 on the closest cooperation of the theorist and the teaching 
 force, largely because the classroom is the laboratory of 
 the experimental educationalist. It is the general belief 
 that the teacher is interested and eager to know what 
 degree of success his efforts have won, as measured by 
 the quality of the work done by his pupils. The diffi- 
 culty which, up to the present time, has confronted the 
 teacher has been the fact that methods for the measure- 
 ment of school products have been so complicated with 
 
 vii
 
 viii Introduction 
 
 statistical data that it has been almost impossible for the 
 ordinary reader to comprehend the movement, and still 
 more impossible for the teacher to apply these methods 
 in the classroom. It is the object of this book to present, 
 in a manner free from statistical data and other compli- 
 cated material, a few of the more important scales which 
 have been worked out and which can profitably be used 
 in the course of ordinary school work by any teacher 
 without special training. No attempt has been made 
 to give a complete presentation of a subject which has 
 been so recently developed, and which is still passing 
 through the experimental stage. The authors have 
 chosen to describe the method of construction of a few of 
 the more important scales rather than to cover all scales 
 which have been published up to the present time. Pro- 
 vided the reader becomes acquainted with the general 
 idea of this scientific movement by using the scales here 
 described, there is no danger that he will fail to use new 
 scales as they are developed. 
 
 Everything has been sacrificed to clearness ; even uni- 
 formity in the method of presentation has had to give 
 way, where such uniformity would have resulted in lack 
 of clearness. 
 
 Each scale has been presented as a unit. This has 
 made necessary a certain amount of repetition, but the 
 advantages of the method are apparent.
 
 THE SCIENTIFIC MEASUREMENT OF 
 CLASSROOM PRODUCTS 
 
 CHAPTER I 
 
 OBJECTIVE VERSUS SUBJECTIVE SCALES OF 
 MEASUREMENT 
 
 More and more, as we come to analyze the educa- 
 tional process, the old idea that this process goes on as 
 a whole is being abandoned. In spite of obvious dangers, 
 the more enlightened view regards the education of any 
 particular individual as the conscious attempt of society 
 to make that individual advance along certain desirable 
 paths — the desirability of any particular path being de- 
 termined by the capacity of the individual and the demands 
 of society. 
 
 Along some lines the school insists that every pupil 
 shall advance; for example, he must improve along the 
 lines of activity which for convenience are termed arith- 
 metic, writing, reading, English composition, spelling, 
 drawing, etc. If we regard the pupil as advancing simul- 
 taneously along all these partly independent lines, we need 
 not, and for many purposes should not, regard education 
 as a single process but rather as a series of processes, each 
 of which, when recognized, admits of study by any per- 
 son who is prepared to take the time to specialize in that 
 direction. Even such aims as the school has in regard to 
 the building of character must, in the same way, be re- 
 garded as an attempt on the part of the school to make 
 the pupils improve on a moral scale, which at present 
 exists merely in thought. 
 
 1
 
 Scientific Measurement 
 
 If those interested in education would consistently take 
 this analytical position, there would be a great change of 
 attitude towards educational problems. For when we 
 look at education as some large general process, the task 
 of improving that process appears rather formidable ; but 
 when it is seen that general improvement is merely ad- 
 vance in certain specific and narrow directions desired by 
 society, the problem of advance becomes, comparatively 
 speaking, a simple one. There still remains the great 
 question of what activity is the most desirable; this 
 question must be left as one of the fundamental problems 
 of the philosophy of education. When, however, it has 
 been decided that a certain line of activity must be pur- 
 sued in the schools, then the question and problem for 
 the ordinary teacher is this : How can I most efficiently 
 train the pupils to improve in skill along this line? 
 What method should I adopt to bring the child to a 
 reasonable efficiency in the minimum time? That is, 
 from the point of view of the teacher there is often no 
 doubt of what has to be done, the problem being, how 
 can the result be accomplished with the greatest economy 
 of effort in the minimum time. In other words, it is not 
 so much a matter of what to teach, as of what method shall 
 be used in teaching. 
 
 But when it comes to a choice between the various 
 methods of teaching a given subject, what is to be the 
 criterion? Is it to be the opinion of the teacher, or of 
 the supervisor, or of the superintendent ? If so, on what 
 is this opinion based ? In education, the time has passed 
 when the inventor of a certain system can make un- 
 challenged claims with regard to the success of his method ; 
 for in many subjects it is now possible to measure, under 
 scientific conditions and independently of any individual 
 judgment, the results obtained under various methods. 
 In this way there is arising a science of method founded 
 on the secure basis of accurate measurement of results.
 
 Objective versus Subjective Scales 
 
 The first essential, then, for the teacher, if sound judg- 
 ment of the success of classroom instruction is to be 
 secured, is to have some scientific methods of measuring, 
 at intervals, the increase of skill or the improvement of 
 the class along the lines of activity for which that teacher 
 is responsible. Just as in a gymnasium it is possible to 
 measure the increase in height from month to month or 
 from year to year by means of a scale of height, so in the 
 school the teacher must be able to measure the rate of 
 improvement of pupils in the various subjects taught. 
 
 This really is no new idea in education ; the school has 
 always been more or less interested in measurement, as 
 the common practice of examinations proves. It is not a 
 question of whether we are to measure the efficiency of the 
 pupils or not, but rather how we are to measure this effi- 
 ciency. Shall the judgment be the opinion — frequently 
 offhand or prejudiced — of some one individual ; or shall 
 the judgment be determined by the use of a standard de- 
 vised by and based upon the consensus of opinion of many 
 experts? If the results of classroom work are to be 
 measured with any degree of exactness, then what we 
 need are scales for measurement, scales which are as inde- 
 pendent as possible of the judgment of the individual who 
 uses them. 
 
 The question, therefore, of urgent importance, is: 
 What are the present methods of measuring efficiency in 
 the schools, and how satisfactory are these methods? It 
 is true that there is perhaps no part of the teacher's work 
 which he knows to be more unsatisfactory than the usual 
 method of awarding grades and marks. Two methods 
 are now in vogue; namely, the percentage method, and 
 the letter mark method. In the former the pupil's work 
 is graded on a basis of 100 as the standard ; in the latter, 
 a letter such as E, G, F, is given to indicate a certain 
 degree of efficiency. 
 
 In a recent article in the Educational Review a well-
 
 Scientific Measurement 
 
 known writer makes the statement: "85% as a class 
 average in subjects like arithmetic or grammar is not 
 excessive." This statement may be true or false, but in 
 any case it is valueless, for the simple reason that a mark 
 of 85% never means the same standard to one individual 
 that it does to another. In a reply to this article one 
 writer states : "What 85% means is absolutely unknown 
 and unknowable — quot homines tot sententiae!" The 
 same argument applies to a grading indicated by letter. 
 What guarantee is there that the same grading represents 
 the same standard of work, when measured by different 
 individuals? All attempts which have been made to in- 
 vestigate this subject prove conclusively that, even in the 
 same school, two teachers will often give the same grading 
 for work which is by no means equivalent. What, then, 
 is wrong with such scales of marking? Obviously, the 
 errors that arise from their use are due to the fact that 
 they depend too much on the individual judgment of the 
 teacher, or, in other words, that the scales are too sub- 
 jective. 
 
 Opinions of teachers on handwriting form an excellent 
 illustration of the dangers and disadvantages of subjective 
 judgments. When a teacher says of a particular sample 
 of writing that it is "good," "fair," "poor," not only does 
 this judgment fail to give an absolute measure of effi- 
 ciency, but even the judgment itself is largely determined 
 by the extent to which the teacher is partial to such 
 characteristics as legibility, grace, character, or to various 
 styles of writing, such as slanting or vertical. In the 
 writing scale to be described, a successful attempt has 
 been made to eliminate this type of unscientific judgment. 
 
 In opposition to these subjective scales of measurement, 
 which depend so much upon the judgment of the individual, 
 there are scales such as those used in measuring mass, 
 length, or time. In the use of these objective scales, 
 very little depends on the judgment of the individual.
 
 Objective versus Subjective Scales 5 
 
 When one says that a particular body weighs 14.6 pounds 
 or that the length of a certain rod is 18.1 feet, there is no 
 room for dispute, since such measurements are outside the 
 range of personal opinion. In other words, they are what 
 we call universal or objective, for they mean the same thing 
 to all persons at all times and in all places. On the con- 
 trary, judgments of plays, books, moral characteristics, 
 depend very much on the character and taste of the in- 
 dividual. The designation "good" used by different 
 individuals may mean very different degrees of merit; 
 that is, the judgments are subjective. In the light of what 
 we have said we may define a perfectly objective scale as 
 a scale in respect to whose meaning all competent thinkers 
 agree ; while a perfectly subjective scale is one in respect 
 to whose meaning all those competent to judge would be 
 likely to disagree, save by chance. 
 
 When subjective scales, such as those described, are 
 used in schools, it is evident that we can have no scien- 
 tific basis for comparison. Yet all agree that improve- 
 ment and advance depend largely upon critical compari- 
 son. Up to the present time, therefore, one of the great 
 methods of obtaining efficiency in the outside world, has 
 not been employed in education, because critical com- 
 parison could not safely be based on subjective judgments. 
 When objective scales are employed in the schools, then 
 it will be feasible to compare the work done in one school 
 with the work done in another, or the work done under 
 one method of instruction with the work done under a 
 different method. Even now, in certain subjects the 
 school administrator is able to compare teacher with 
 teacher, school with school, system with system, and even 
 country with country. 
 
 The great problem of measurement in education, there- 
 fore, is to construct objective or universal scales, about 
 the use of which there can be no misunderstanding when 
 they are placed in the hands of competent teachers.
 
 6 Scientific Measurement 
 
 Every such scale must fulfill at least three essential re- 
 quirements: (1) It must measure a desired product; 
 (2) it must be so simple in its application that it is suit- 
 able for ordinary classroom use ; (3) it must not require 
 an undue amount of time in administration. 
 
 From the very nature of measurement it is apparent 
 that ability in such a subject as arithmetic admits of being 
 measured objectively. Any competent teacher would be 
 capable of constructing a scale to measure improvement 
 in addition. But the essential thing is that every one 
 shall agree to use the same method, or standard, just as 
 they agree to use a gram, a centimeter, and a second in 
 measuring mass, length, and time. Thus, suppose it is 
 desired to measure speed in adding, all that is necessary 
 is to construct a blank on which are printed the columns 
 of figures. The test can then be administered by allow- 
 ing, let us say, two minutes for the work, that is, less 
 time than it takes even the fastest pupil to complete all 
 the addition. Provided the same directions are followed 
 in each case, it is possible to measure by the same stand- 
 ard any other school in any other system. In this way a 
 comparison of the two groups will be perfectly easy. The 
 essential point, then, is that all shall agree to use the same 
 scale, under the same conditions, giving the same time 
 allowance, and correcting and scoring in the same way. 
 It is to fulfill these conditions that objective scales are 
 necessary. 
 
 While arithmetic lends itself to such objective measure- 
 ments, in other important subjects it is more difficult to 
 construct scales for the measurement of efficiency, which 
 will be relatively independent of the judgment of the 
 teacher. It would be ideal if scales could be constructed 
 which would measure improvement in writing, reading, 
 drawing, English composition, spelling, etc., about the 
 use of which there would be as little division of opinion 
 as there is about the employment of a yardstick, a balance,
 
 Objective versus Subjective Scales 7 
 
 or a watch, to measure length, mass, and time, respec- 
 tively. In the following pages a few of the more essential 
 objective scales which have been worked out with this 
 idea in view, will be presented. No claim is made that 
 they eliminate completely the factor of the judgment of 
 the individual teacher. Through the description and use 
 of the scales themselves the reader may judge of the extent 
 to which individual opinion, bias, and prejudice, as fac- 
 tors, have been excluded. 
 
 It will be seen that a scale may be used by the teacher 
 merely to measure the improvement of a particular class 
 or individual. It is advantageous, however, after a par- 
 ticular test has been administered, to know how the grade 
 taking the test compares with similar grades in other 
 school systems. In certain cases it is possible to make 
 this comparison, for the scales have been tested with a 
 sufficient number of pupils to establish averages of achieve- 
 ment, or, in other words, norms or standards for the various 
 grades. The process of standardizing a test is quite 
 simple. All that is necessary is to administer the test, 
 under the identical standard conditions, in like grades, in 
 different representative school systems, and from these 
 results to determine the average work done in the various 
 grades. It will be noted, from what follows, that we do 
 not have a different scale for each grade, for in many cases 
 all grades can be measured by the same scale. Just as 
 we measure the dwarf and the giant with the same foot 
 rule, and express the result in the same unit, inches, so 
 we may measure the ability of individuals at different 
 points of their training on the same scale, expecting of 
 course increasing products at successive stages. In so far 
 as standards have been established for the grades, they 
 are included in the description given in the following chap- 
 ters (II-VIII). In all cases these standards or norms must 
 be looked upon as provisional, for none of the tests have 
 been tried upon a sufficiently wide range of schools and
 
 8 Scientific Measurement 
 
 school systems to make it certain that the standards are 
 the average achievements of the particular grade in 
 question. 
 
 EXERCISES 
 
 1. What are the present methods by which you measure the effi- 
 ciency of your class work? Why is it so difficult to tell how your 
 class compares with other classes of the same grade? How would 
 such information increase your efficiency? 
 
 2. Taking twenty compositions of varying merit, grade them 
 according to your usual method. Put them away for a month and 
 then grade again. How do the results compare ? Repeat the experi- 
 ment with a series of handwriting samples. 
 
 3. Does the same mark given by different teachers imply the same 
 standard of work on the part of the pupils ? How would you prove 
 the correctness of your answer? 
 
 4. What is the final test of any particular method of teaching a 
 subject? Why is there so much difference of opinion with reference 
 to the relative values of different methods? 
 
 5. What would happen if a foot meant a different length in dif- 
 ferent parts of the country? What is the effect of 80% in one school 
 not meaning the same as 80% in another school? 
 
 6. Give ten examples of qualities which we measure (a) objec- 
 tively; (b) subjectively. Is there any such thing as a perfect 
 objective measure? Give your reasons. 
 
 7. What is meant by a norm ? How would you establish the norms 
 of height and weight of the grades in a school? How could we use 
 the same idea in measuring growth in spelling, arithmetic, writing, 
 and reading skill? What are the difficulties? 
 
 8. What are the disadvantages of the present system of subjective 
 marking as they affect (a) the pupil ; (b) the teacher ; (c) the adminis- 
 tration of the school system? 
 
 9. It has been shown that the same arithmetic paper received 
 85% and 40% when marked by two trained judges; how could 
 this happen? What might have been done to avoid it? 
 
 10. If you had 5000 arithmetic papers, and five judges had each 
 to mark 1000 of these, how would you attempt to secure uniformity 
 in the system of marking?
 
 CHAPTER II 
 ARITHMETIC SCALES 
 
 I. COURTIS TESTS IN ARITHMETIC 
 H. WOODY TESTS IN ARITHMETIC 
 
 When it is considered how many different operations 
 are covered by the inclusive term "arithmetic," it be- 
 comes apparent at once how little specific meaning is con- 
 veyed by the assertion that a pupil is good or poor in that 
 subject. Since arithmetical skill is not a single ability, 
 but consists rather of a number of abilities, discussion of 
 it should be expressed in terms of these. For instance, 
 instead of saying that a child is good in arithmetic, it is 
 more accurate and far more useful to state in what specific 
 process or processes — adding, subtracting, reasoning, etc. 
 — he excels, for a child may be good in one of these opera- 
 tions and poor in another. Thus, the teacher is con- 
 fronted with a problem of analysis. It must be discovered 
 first of all in what particular process— adding, multiplying, 
 subtracting, etc. — the pupil is weak, and then an attempt 
 must be made to strengthen him in that process. To facili- 
 tate this work is the chief object of the Courtis Tests in 
 Arithmetic. 
 
 Eight or nine years ago, in an early effort to meas- 
 ure efficiency in certain phases of arithmetical work, 
 Courtis discovered that the ability of a given individual 
 in some one process was very different from his ability 
 in another. One child might be very good in addition 
 and poor in multiplication, while another might be good 
 in both addition and multiplication but poor in reasoning, 
 etc. Courtis immediately began experimental work to 
 control this individual variation; that is, to make the 
 
 9
 
 10 Scientific Measurement 
 
 child do equally well in all these operations. For several 
 years the attempt failed ; for with his increased effort at 
 control, Courtis found that the difference in the ability 
 of the individual in the various branches also increased. 
 However, this work was not without most important re- 
 sults. As an outcome of administering the tests to more 
 than 48,000 children in about 70 schools in 10 states, 
 Courtis discovered one fundamental fact; namely, that 
 one of the great factors in education is the variability of 
 the natural abilities of children. In the first place, no 
 child will do equally well in all the operations involved in 
 arithmetic ; for example, he may do very well in division 
 and still do poorly in addition, or vice versa. That is, 
 there is a difference among his special attainments or 
 abilities in these sub-branches. Secondly, there exists a 
 great difference in the general ability of different children. 
 According to Courtis, these two facts mean that new edu- 
 cational methods, methods that will give each child a 
 chance to develop in his own way and along his own lines, 
 will have to be invented. In the work of analysis thus 
 necessitated, the Courtis Tests will be of great assistance.
 
 Arithmetic Scales 
 
 11 
 
 I. COURTIS TESTS 
 
 These original tests, called by Courtis " Series A," are 
 eight in number and are designed to test those abilities 
 which constitute most of that complex product known as 
 arithmetical efficiency. 
 
 
 Courtis Tests — Series A 
 
 
 Number 
 
 Function Time for Administra- 
 
 of Test 
 
 
 TION OF TE8T8 
 
 1 
 
 Addition 
 
 
 One minute 
 
 2 
 
 Subtraction 
 
 Combinations 
 
 a tt 
 
 3 
 
 Multiplication 
 
 0-9 
 
 << << 
 
 4 
 
 Division 
 
 
 << << 
 
 5 
 
 Copying Figures 
 (Rate of motor activity) 
 
 « << 
 
 6 
 
 Speed Reasoning 
 (Judgments of operation to be used in 
 simple one-step problems) 
 
 << << 
 
 7 
 
 Fundamentals 
 (Abstract examples in the four opera- 
 tions) 
 
 Twelve minutes 
 
 8 
 
 Reasoning 
 
 Six 
 
 
 (Two-step pro 
 
 blems) 
 
 
 These eight tests are printed on separate sheets of paper, 
 and folders, containing full directions intended to secure 
 uniformity of administration and marking, may be had 
 by the examiner. For example, in the Addition Test 
 given on page 12, the child is supposed to add across the 
 paper from left to right. That is, his answers should be, 
 9, 18, 13, 8, etc. He should do as many of these prob- 
 lems as possible in the time allowed — one minute — 
 and his score will be the number of problems he has done 
 correctly in that time.
 
 12 Scientific Measurement 
 
 Arithmetic — Test No. i Speed Test — Addition 
 
 Write on this paper, in the space between the lines, 
 the answers to as many of these addition examples as 
 possible in the time allowed. 
 
 89782 13603 17932 16904 
 19605 58972 37604 26512 
 
 58694 12567 34703 14802 
 13503 49802 16985 67957 
 
 18605 48953 13823 29745 
 94724 18706 79507 23802 
 
 37904 24516 92506 74803 
 34865 18902 18743 19604 
 
 69812 16702 59675 48507 
 
 14713 84536 52803 42693 
 
 i 
 
 14904 12603 89785 17932 
 
 67512 67972 19602 37604 
 
 (Copyright by S. A. Courtis)
 
 Arithmetic Scales 13 
 
 The same mode of procedure applies to the Subtraction, 
 Multiplication, and Division Tests. 
 
 In the test of reasoning with one-step problems (Test 
 No. 6), the pupil is not required to work out the prob- 
 lems but merely to record what operation — addition, 
 subtraction, etc. — he would use if he were going to work 
 them out; this is to distinguish between skill in reason- 
 ing and mere skill in rapid calculation. In the reasoning 
 test involving two steps (Test No. 8), the answer is to be 
 found and recorded. 
 
 The tests just described were designed to measure the 
 relation existing between the simpler abilities tested in the 
 first six tests and the more complex abilities tested in 
 the last two tests. That is, their object was to investigate 
 whether a child who is good or poor in addition, subtrac- 
 tion, multiplication, or division is also good or poor, as 
 the case may be, in fundamentals and reasoning. Courtis 
 claims the tests have accomplished this purpose.
 
 14 
 
 Scientific Measurement 
 
 Arithmetic — Test No. 7 — Fundamentals 
 
 In the blank space below, work as many of these ex- 
 amples as possible in the time allowed. Work them in 
 order as numbered, writing each answer in the "answer 
 column" before commencing a new example. Do not 
 work on any other paper. 
 
 No. Operation 
 
 1 Addition 
 
 Example Answer 
 
 a. 25 + 830 + 122 = 
 
 b. 232 + 8021 + 703 + 3030 = 
 
 2 Subtraction 
 
 a. 5496 - 163 = 
 
 b. 943276 - 812102] = 
 
 3 Multiplication 
 
 4 Division 
 
 5 Addition 
 
 6 Subtraction 
 
 g I Multiplication 
 
 9 
 
 10 
 11 
 
 Division 
 Division 
 
 \ 2 3 ] Addition 
 
 14 Subtraction 
 j^} Multiplication 
 17 Division 
 !g} Division 
 
 2012 X 213 = 
 
 158664 -T- 132 = 
 
 6134 + 213 + 4800 + 6005 + 474 
 
 73210142 - 49676378 = 
 
 46505 X 456 = 
 
 27217182 
 
 6 = 
 
 3127102 -=- 463 = 
 85586 + 69685 + 39397 + 
 95836 + 37768 + 69666 + 
 78888 + 54987 = 
 15655431 - 5878675 = 
 
 78965 X 678 = 
 
 44502486 -=- 7 = 
 
 5373003 + 769 = 
 
 (Copyright by S. A. Courtis.)
 
 Arithmetic Scales 15 
 
 Arithmetic — Test No. 8 — Reasoning 
 
 In the blank space below, work as many of the following 
 examples as possible in the time allowed. Work them in 
 order as numbered, entering each answer in the " answer 
 column " before commencing a new example. Do not 
 work on any other paper. 
 
 1. A party of children went from a school to a woods to gather 
 nuts. The number found was but 205, so they bought 1,955 nuts 
 more from a farmer. The nuts were shared equally by the children 
 and each received 45. How many children were there in the party? 
 
 2. One summer a farmer hired 43 boys to work in an apple orchard. 
 There were 35 trees loaded with fruit and in 57 minutes each boy had 
 picked 49 apples. If in the beginning the total number of apples on 
 the trees was 19,677, how many were there still to be picked? 
 
 3. A girl found by careful counting that there were 87 letters more 
 on a page in her history than on a page of her reader. She read 31 
 pages in each book in the first 29 days of school. How many more 
 letters each day did she read in one book than in the other ? 
 
 4. The children of a school made small boxes to be filled with 
 candy and given as presents at a school party. Six hundred were 
 needed. In 4 days grades III to VII made 20, 25, 83, 150 and 150 
 boxes. The eighth grade agreed to make the rest. How many did the 
 eighth grade make? 
 
 5. A girl's record in spelling for 5 days was 19, 18, 20, 16 and 20 
 words spelled correctly out of 20. If each of the 16 children in the 
 grade had had the same record, what would have been the total num- 
 ber of words spelled correctly by that grade in 5 days? 
 
 6. A party of boys went on a long bicycle trip. They traveled 
 1702 miles in 37 days. A number of men then joined the party, and 
 soon the party was traveling 58 miles per day. How much change 
 in the number of miles ridden a day did the presence of the men make? 
 
 7. A teacher corrected 2400 arithmetic test papers ; 2295 of these 
 he marked "poor," "good" etc. All the others were marked "un- 
 satisfactory." If each of the papers in this group had 47 mistakes, 
 what was the total number of mistakes in the unsatisfactory papers? 
 
 8. In two schools five teachers recorded the number of blocks the 
 
 children walked in going to and from the school. The total for one 
 
 school was 3000 blocks ; for the other 2400. The number of children 
 
 in both schools was 216. How many blocks did each child walk a 
 
 day' 
 
 (Copyright byS. A. Courtis)
 
 16 Scientific Measurement 
 
 Courtis Tests — Series B 
 
 Courtis has recently constructed a new series of more 
 difficult tests, "Series B," to be used in the primary grades 
 for testing more complex operations in the four funda- 
 mental processes. The figures in these tests are chosen so 
 that all the fundamental combinations are included. 
 
 Number op Function Time for Administra- 
 
 Test tion of Tests 
 
 1 Addition Eight minutes 
 
 2 Subtraction Four 
 
 3 Multiplication Six 
 
 4 Division Eight 
 
 In the Addition Test the pupil is required to add as 
 many figures as possible in eight minutes. In this way it 
 may be determined whether or not a child or class has 
 learned (1) the fundamental combinations ; (2) the mechan- 
 ism of column addition; (3) to carry; (4) to hold the 
 attention ; (5) to control the effects of fatigue or bore- 
 dom; (6) to work at a high speed; (7) to work with 
 accuracy. In a similar manner, each of the other three 
 tests is put in the simplest form necessary to serve as a 
 general measure of ability in that operation. 
 
 Test No. i — Addition 
 
 You will be given eight minutes to find the answers to 
 as many of these addition examples as possible. 
 
 927 
 
 297 
 
 136 
 
 486 
 
 384 
 
 176 
 
 277 
 
 837 
 
 379 
 
 925 
 
 340 
 
 765 
 
 477 
 
 783 
 
 445 
 
 882 
 
 756 
 
 473 
 
 988 
 
 524 
 
 881 
 
 697 
 
 682 
 
 959 
 
 837 
 
 983 
 
 386 
 
 140 
 
 266 
 
 200 
 
 594 
 
 603 
 
 924 
 
 315 
 
 353 
 
 812 
 
 679 
 
 366 
 
 481 
 
 118 
 
 110 
 
 661 
 
 904 
 
 466 
 
 241 
 
 851 
 
 778 
 
 781 
 
 854 
 
 794 
 
 547 
 
 355 
 
 796 
 
 535 
 
 849 
 
 756 
 
 965 
 
 177 
 
 192 
 
 834 
 
 850 
 
 323 
 
 157 
 
 222 
 
 344 
 
 124 
 
 439 
 
 567 
 
 733 
 
 229 
 
 953 
 
 525 
 
 (Copyright by S. A. Courtis)
 
 Arithmetic Scales 17 
 
 Series B may be used from the fourth grade up. When 
 these four tests are standardized, which will take place 
 as soon as more returns from their use are available, it 
 will be possible to tell the degree of skill in each test 
 which the average child in any particular grade should 
 attain. 
 
 It is to be remembered that the Courtis Tests are 
 "neither lesson sheets nor examination papers." They 
 are only methods of investigation — mere measuring rods. 
 By their use are revealed the actual arithmetical condi- 
 tions existing in schools, classes, and individuals. To 
 find the causes of any unsatisfactory conditions which 
 the tests may reveal, and to remove these causes, is an- 
 other problem. 
 
 The repeated use of these scales will tend to reveal the 
 laws of development as they operate in the classroom, 
 and to measure the efficiency of any particular educational 
 method. A class which is being taught division by a 
 certain method may be tested at intervals to see what 
 improvement is taking place. If little or no improvement 
 is shown, it may be safely inferred that the method used 
 is not suited to that particular group ; and new methods 
 of instruction may be devised and tested, until the im- 
 provement is so marked as to leave little doubt that 
 the method finally adopted is the one which produces the 
 best results with the group in question. In short, the 
 tests of Series B are scientific measures of efficiency in four 
 operations of arithmetic, which may be used to determine 
 the best methods of teaching these operations. Since 
 the same tests, or their equivalents, are used in all the 
 grades, a child or group of children may be measured 
 over and over again, and the progress determined by the 
 changes in the score, just as height is measured over and 
 over again with the same measuring rod. 
 
 Since the use of standard tests makes objective scoring 
 possible, any teacher can easily establish objective stand-
 
 18 Scientific Measurement 
 
 ards of work for a class ; and in time it will be known 
 what the actual standards for different school systems 
 are. To facilitate this work of standardization, Courtis 
 has published printed folders of instruction covering every 
 phase of the testing, such as scoring, tabulating results, 
 the making of graphs, etc. Very likely Series B will 
 eventually displace Series A, except for the solution of 
 special problems, and standards of permanent value will 
 be obtained from its use. To those who wish to give a 
 single test, merely to see the nature of the experiment or 
 to measure the general character of the arithmetical work 
 of a class as compared with that done in another class or 
 school, the test on fundamentals (Test No. 7) in Series A 
 is recommended, for it is a general measure of the ability 
 to add, subtract, multiply and divide with whole 
 numbers. 
 
 The administration of these tests is an easy matter. 
 The twelve tests — the eight of Series A and the four of 
 Series B — are printed on separate sheets of paper, each 
 containing complete directions for its use. It is advisable 
 to procure with these test sheets the manual containing 
 full directions for the giving of the tests ; for the essence 
 of this movement lies in uniformity of administration 
 and marking. The Courtis Standard Tests for Arith- 
 metic may be obtained at the Department of Cooperative 
 Research, 82 Eliot Street, Detroit, Michigan. 
 
 If a teacher desires merely to compare the general char- 
 acter of the work of a class with the work of other classes 
 of the same grade, all that is necessary is to send for 
 Test No. 7, Series A, together with the folder relating to 
 the tests of that series. If, after the administration of this 
 single test, more specific information is desired regarding 
 the work of pupils in the various sub-branches— addition, 
 subtraction, etc. — other tests in Series A, Test No. 1, 
 for addition, Test No. 2, for subtraction, and so on, may 
 be procured and administered. In the fourth grade and
 
 Arithmetic Scales 19 
 
 above it is advisable to use the tests of Series B, as they 
 are cheaper and require less time to administer. 
 
 The actual application of the test is very simple and 
 requires but little time — from one to twelve minutes 
 according to the test. For example, in giving Test No. 1, 
 Series A (Addition), the teacher, after reading the instruc- 
 tions for administering the test as given in the manual 
 for Series A, will proceed somewhat as follows. Holding 
 up one of the test sheets before the pupils, the teacher 
 will give directions for filling out the blank spaces at the 
 top of the paper with the name of pupil, the grade, and 
 name of school. Then, in a manner calculated to secure 
 cooperation, the pupils will be told just what is expected 
 of them; namely, at a given signal, "Start," to add 
 across the paper from left to right, putting down the 
 answers in the spaces allowed between the lines until 
 the signal, "Stop," is given. This signal is given after 
 one minute's time has elapsed. The teacher later records 
 the number of problems each child has done correctly. 
 This constitutes his score. A more or less similar course 
 is followed with each of the other tests. 
 
 The only warning to be observed in the administration 
 of the tests is that care must be taken to see that all the 
 pupils start and stop at the same time, and that every 
 effort be made to secure the interest and cooperation of 
 the children. The work itself should proceed smoothly 
 and steadily with no hurry or excitement. Class averages 
 may be obtained from the record of the individual scores 
 and such averages may be compared with those obtained 
 from different parts of the country. (See Standard 
 Scores.) 
 
 Within the classroom the teacher is in a position to 
 determine which children should be selected for special 
 attention. For example, if a child's record shows him 
 to be very high in multiplication and low in addition, 
 efforts should be made to improve the latter, and he
 
 20 
 
 Scientific Measurement 
 
 should not be made to waste time on multiplication drills. 
 Tests administered at the beginning of school in Septem- 
 ber will show what children fall below the standard for 
 each process in that grade. Several tests during the 
 year will show the efficiency or inefficiency of the methods 
 used to bring these children's records up to standard. 
 Furthermore, a child's improvement may be followed 
 from grade to grade by keeping a record of each pupil's 
 score. The results obtained from the administration of 
 such tests also make possible the accurate comparison 
 of school systems and classes. These tests mean better 
 work on the part of teachers because they reveal just 
 what they are accomplishing; they mean progressive 
 educational changes brought about through those methods 
 of instruction which have produced the best results. 
 
 STANDARD SCORES 
 
 As a result of administering the eight tests in Series A 
 to almost 6700 pupils throughout the United States, 
 Courtis has worked out the following tentative standard 
 scores. These, it should be noted, are the average scores 
 actually obtained by the pupils themselves. 
 
 
 l 
 
 2 
 
 3>4 
 
 5 
 
 No. 6 
 
 No. 7 
 
 No. 8 
 
 Test No. 
 
 Atts. 
 
 Rts. 
 
 Atts. 
 
 Rts. 
 
 Atts. 
 
 Rts. 
 
 Grade III . . . 
 
 26 
 
 19 
 
 16 
 
 58 
 
 2.7 
 
 2.1 
 
 5.0 
 
 2.7 
 
 2.0 
 
 1.1 
 
 Grade IV . . . 
 
 34 
 
 25 
 
 23 
 
 72 
 
 3.7 
 
 3.0 
 
 7.0 
 
 3.3 
 
 2.6 
 
 1.7 
 
 Grade V . . . 
 
 42 
 
 31 
 
 30 
 
 86 
 
 4.8 
 
 4.0 
 
 9.0 
 
 4.9 
 
 3.1 
 
 2.2 
 
 Grade VI . . . 
 
 50 
 
 38 
 
 37 
 
 99 
 
 5.8 
 
 5.0 
 
 11.0 
 
 6.6 
 
 3.7 
 
 2.8 
 
 Grade VII . . . 
 
 58 
 
 44 
 
 44 
 
 110 
 
 6.8 
 
 6.0 
 
 13.0 
 
 8.3 
 
 4.2 
 
 3.4 
 
 Grade VIII . . 
 
 63 
 
 49 
 
 49 
 
 117 
 
 7.8 
 
 7.0 
 
 14.0 
 
 10.0 
 
 4.8 
 
 4.0 
 
 Grade IX . . . 
 
 65 
 
 50 
 
 50 
 
 120 
 
 8.6 
 
 7.8 
 
 15.0 
 
 11.0 
 
 5.0 
 
 4.3 
 
 Time allowances, 
 
 
 
 
 
 
 
 
 
 
 
 minutes . . . 
 
 1 
 
 1 
 
 1 
 
 1 
 
 6 
 
 6 
 
 12 
 
 12 
 
 6 
 
 6
 
 Arithmetic Scales 21 
 
 Thus in the Addition Test (Test No. 1), the average 
 score in Grade V is 42, the number of correct additions 
 made in one minute. Similarly, for all the other tests. 
 
 II. WOODY ARITHMETIC SCALE 1 
 
 Whereas, in each of the separate Courtis Tests the 
 problems are of approximately the same difficulty through- 
 out, in the Woody Scales a different method of measuring 
 efficiency is employed. The scales are designed to measure 
 work in the four fundamental operations of (a) addition, 
 (6) subtraction, (c) multiplication, and (d) division, re- 
 spectively. Each of these scales consists of a great variety 
 of problems falling within the field of the particular opera- 
 tion that the scale is designed to test. These problems, 
 beginning with the easiest that can be found, gradually 
 increase in difficulty until the last ones in each scale are 
 so difficult that only a relatively small percentage of the 
 pupils in the eighth grade are able to solve them correctly. 
 That is, taking the addition scale for example, the problems 
 rise in difficulty from the first, which requires next to no 
 ability in addition, up to the last, which, though still an 
 addition problem, is of sufficient complexity to test chil- 
 dren of the eighth grade. The relative difficulties of the 
 problems within each scale were determined by adminis- 
 tering them to large groups of children in several school 
 systems, the difficulty of a problem being calculated 
 from the percentage of correct answers by a method 
 similar to that used in the Buckingham Spelling Scale. 
 
 Two distinct series of scales in each of the above named 
 operations have been devised. It will be sufficient here to 
 describe the shorter of these scales, Series B, and to illus- 
 trate the general principles which underlie this method of 
 measurement. For the other scale with a full account of 
 its instructions, method of administration, scoring, etc., 
 the reader is referred to the original study. 
 
 1 The scales are reproduced by the courtesy of Dr. Clifford Woody.
 
 22 
 
 Scientific Measurement 
 
 Series B — Addition Scale 
 
 Name 
 
 When is your next birthday? How old will you be? 
 
 Are you a boy or girl ? In what grade are you ? . . . 
 
 (1) 
 
 (2) 
 
 (3) 
 
 (5) 
 
 (7) 
 
 (10) 
 
 2 
 
 2 
 
 17 
 
 72 
 
 3 + 1 = 
 
 21 
 
 3 
 
 4 
 
 2 
 
 26 
 
 
 33 
 
 
 3 
 
 — 
 
 
 
 
 35 
 
 (13) 
 
 (14) 
 
 
 (16) 
 
 (19) 
 
 (20) 
 
 23 
 
 25 +42 
 
 = 
 
 9 
 
 $.75 
 
 $12.50 
 
 25 
 
 
 
 24 
 
 1.25 
 
 16.75 
 
 16 
 
 
 
 12 
 
 .49 
 
 15.75 
 
 
 
 
 15 
 19 
 
 
 
 (21) 
 
 (22) 
 
 
 (23) 
 
 (24) 
 
 (30) 
 
 $8.00 
 
 547 
 
 
 i + i = 
 
 4.0125 
 
 2§ 
 
 5.75 
 
 197 
 
 
 
 1.5907 
 
 61 
 
 2.33 
 
 685 
 
 
 
 4.10 
 
 3| 
 
 4.16 
 .94 
 
 678 
 456 
 
 
 
 8.673 
 
 — 
 
 
 
 
 
 6.32 
 
 393 
 525 
 240 
 152 
 
 
 
 
 
 (33) 
 
 (36) 
 
 
 
 (38) 
 
 
 .49 
 
 2 yr. 5 mo. 
 
 25.09 
 
 .28 
 
 3 yr. 6 mo. 
 
 
 
 
 
 .63 
 
 4 yr. 9 mo. 
 
 
 
 
 
 .95 
 
 5 yr. 2 mo. 
 
 
 
 
 
 1.69 
 
 6 yr. 7 mo. 
 
 
 
 
 
 .33 
 .36 
 
 1.01 
 .56 
 .88 
 .75 
 .56 
 
 1.10 
 .18 
 .56
 
 Arithmetic Scales 23 
 
 Series B — Subtraction Scale 
 
 Name 
 
 When is your next birthday? How old will you be?. 
 
 Are you a boy or girl ? In what grade are you ? 
 
 (J) (3) (6) (7) 
 
 8 2 11 13 
 
 5 17 8 
 
 (9) (13) (14) (17) 
 
 78 16 50 393 
 
 37 9 25 178 
 
 (19) (20) (24) (25) 
 
 567482 2| - 1 - 81 27 
 
 106493 5f 12| 
 
 (27) (31) (35) 
 
 5 yds. 1 ft. 4 in. 7.3 - 3.00081 - 31 - If = 
 
 2 yds. 2 ft. 8 in. 
 
 Series B — Multiplication Scale 
 
 Name 
 
 When is your next birthday? How old will you be? 
 
 Are you a boy or girl? In what grade are you? 
 
 (1) (3) (4) (5) 
 
 3X7= 2X8- 4X8= 23 
 
 3 
 
 (8) (9) (11) (12) 
 
 50 254 1036 5096 
 
 8 6 8 6 
 
 (13) (16) (18) (20) 
 
 8754 7898 24 287 
 
 8 9 234 .05 
 
 (24) (26) (27) (29) 
 
 16 9742 6.25 * X 2 
 
 2f 59 3.2 
 
 (33) (35) (37) (38) 
 
 2* X 8* «= 987* 2i X 4i X U - .0963* 
 25 .084
 
 24 Scientific Measurement 
 
 Series B — Division Scale 
 
 Name 
 
 When is your next birthday? How old will you be?. 
 
 Are you a boy or girl ? In what grade are you ? 
 
 (1) 
 
 3j6" 
 
 (2) 
 9J27 
 
 (7) 
 
 4 4-2 = 
 
 (8) 
 9JU 
 
 (11) 
 2JIF 
 
 (14) 
 8)5856 
 
 (15) 
 
 \ of 128 = 
 
 (17) 
 
 50 4-7 = 
 
 (19) 
 
 248 4- 7 = 
 
 (23) 
 23)469 
 
 (27) 
 
 I of 624 = 
 
 (28) 
 .003). 0936 
 
 (30) 
 §4-5 = 
 
 (34) 
 62.50 - 
 
 -U = 
 
 (36) 
 
 9)69 lbs. 9 oz. 
 
 Series B was especially constructed for use in the 
 measurement of arithmetical ability when the amount 
 of time for such measurement is limited. The break 
 in continuity in the numbering of the problems does not 
 mean that the whole scale is not presented. The scale 
 is quite complete as it stands ; the numbering is a matter 
 of convenience for purposes external to the use of the scale. 
 
 The Addition and Subtraction Scales can be used in 
 Grades II to VIII inclusive; the Multiplication and 
 Division Scales, in Grades III to VIII inclusive. It is 
 recommended that in the use of Series B all tests be given 
 together. 
 
 DIRECTIONS FOR ADMINISTRATION 
 
 It is very necessary that the same standard method be 
 employed in the giving of these tests; care should be 
 taken that the same directions are given in the same way 
 to all groups taking the tests. The following are the 
 general directions which should be carefully followed: 
 Distribute the papers face down and do not allow the 
 pupils to turn them over until they are told to do so. 
 When all are ready with pencils in hand, say : " Turn 
 your papers over and answer the questions at the top
 
 Arithmetic Scales 25 
 
 of the page." When all these preliminary questions 
 have been answered, repeat the following formula of 
 specific directions. If you are giving the Addition test, 
 say, " Every problem on the sheet which I have given 
 you is an addition problem, an ' and problem.' Work as 
 many of these problems as you can and be sure that you 
 get them right. Do all your work on this sheet of paper 
 and don't ask anybody any questions. Begin." 
 
 For each test in Series B allow ten minutes. It is 
 essential that all the pupils start and stop work together 
 because the test is partly one of speed. Most of the 
 children will have finished all that are within range of 
 their ability before the end of the time allowed; those 
 who have not must not be allowed any further time. 
 
 The only variation in procedure in giving any of the 
 other tests is the substitution in the formula of specific 
 directions of the expressions "subtraction or 'take away 
 problems,'" "multiplication or 'times problems,'" and 
 " division or ' into problems,' " for the expression " addi- 
 tion or 'and problems.'" Since teachers in the lower 
 grades sometimes use the expressions " and," " take away," 
 " times," and " into," problems, these forms should also 
 be used in administering the test so as to make clear to 
 the children what is expected of them. 
 
 DIRECTIONS FOR SCORING THE TESTS 
 
 In scoring each test the standard of marking should be 
 absolute accuracy and the final answer should be in its 
 lowest terms. 
 
 If the results of class measurement are to be compared 
 with the results and values established by the author, 
 only those answers should be accepted as correct which 
 are identical with those given in the following table, since 
 these are the solutions upon the basis of which the original 
 scoring was done.
 
 26 
 
 Scientific Measurement 
 
 Answers to Problems. Series B 
 
 Addition 
 
 Subtraction 
 
 Problem 
 
 Answer 
 
 Problem 
 
 Answer 
 
 1 
 
 2 
 
 3 
 
 5 
 
 7 
 
 10 
 
 13 
 
 14 
 
 16 
 
 19 
 
 20 .... . 
 
 21 
 
 22 
 
 23 
 
 24 
 
 30 
 
 33 
 
 36 
 
 38 
 
 5 
 
 9 
 19 
 
 98 
 
 4 
 
 89 
 
 64 
 
 67 
 
 79 
 
 $2.49 
 
 $45.00 
 
 $27.50 
 
 3,873 
 
 a 
 
 18.3762 
 
 12| not 11 V s = 
 1 * 
 
 ■•-8 
 
 10.55 
 22 yrs. 5 mo. or 
 
 22 r \ yrs. 
 268.1324 
 
 1 . . 
 
 3 . . 
 
 6 . . 
 
 7 . . 
 9 . 
 
 13 . 
 
 14 . 
 17 . 
 
 19 . 
 
 20 . 
 
 24 . 
 
 25 . 
 27 . 
 
 31 . 
 35 . 
 
 
 
 3 
 1 
 
 4 
 5 
 
 41 
 
 7 
 
 25 
 
 215 
 
 460,989 
 
 If 
 
 3| 
 
 14 a 
 
 2 yds. 1 ft. 8 in. 
 
 not 81in. 
 4.29919 
 2| not 2| = J 
 
 Multiplication 
 
 Division 
 
 Problem 
 
 Answer 
 
 Problem 
 
 Answer 
 
 1 
 
 3 
 
 4 
 
 5 
 
 8 
 
 9 
 
 11 
 
 12 
 
 13 
 
 16 
 
 18 
 
 20 
 
 24 
 
 26 
 
 27 
 
 29 
 
 33 .... . 
 35 
 
 Off n • • • • 
 
 38 
 
 21 
 6 
 
 32 
 
 69 
 
 150 
 
 1,524 
 
 8,288 
 
 30,576 
 
 70,032 
 
 71,082 
 
 5,616 
 
 14.35 
 
 42 
 
 574,778 
 
 20,000 
 
 si 
 
 24693f 
 .0080902^ or 
 
 1 . . 
 
 2 . . 
 
 7 . . 
 
 8 . . 
 11 . . 
 
 14 . 
 
 15 . 
 17 . 
 19 . 
 23 . 
 
 27 . 
 
 28 . 
 30 . 
 34 . 
 36 . 
 
 
 
 2 
 3 
 2 
 
 
 6£ not 6 + 1 
 732 
 
 32 
 
 7| not 7 + 1 
 35f not 35+3 
 20ft ; 20.3, not 
 
 20+9 
 546 
 31.2 
 & or .15 
 
 50 
 7* lbs. 11 § oz.; 
 
 7^1bs. 9 oz. 
 
 
 .00809025 
 
 
 

 
 Arithmetic Scales 27 
 
 METHOD OF DETERMINING THE CLASS ACHIEVEMENT 
 
 The method used for determining the class achieve- 
 ment with Scale B is simpler than that employed in the 
 use of Scale A. It is largely for this reason that Scale B 
 was chosen for description. It should be noted that in 
 each of the scales a definite attempt has been made to 
 place the problems so that they would increase by uni- 
 form stages of difficulty from the first to the last. Thus, 
 in the Addition Scale problem 3 is as much more difficult 
 than problem 2 as problem 5 is more difficult than problem 
 3, and so on. If one compares this with the method of 
 the Courtis Tests, it will be seen that in the latter the 
 problems involving a given operation are all of approxi- 
 mately the same difficulty and require precisely the same 
 knowledge and method for their solution. In other words 
 the Courtis Tests measure speed in the various operations 
 in arithmetic rather than extent of knowledge of the 
 operation involved. In the Woody Scale, because the 
 problems increase in difficulty, the score measures a 
 certain extent of knowledge of the process involved in 
 the operation rather than mere speed of performance. 
 For example, in a race one could have a series of hurdles 
 of all the same height and test the number cleared in a 
 certain time — such is the Courtis method ; or the 
 hurdles could get gradually higher and higher, the success of 
 the individual being measured by the hurdles he can clear 
 without a fall — such is the method of the Woody Scales. 
 
 An objection is sometimes made by teachers that the 
 problems are too hard for the children. In this con- 
 nection it cannot be pointed out too clearly that when 
 scales of this type are used in the schools it is not ex- 
 pected that the children will be able to do all the prob- 
 lems, just as when we determine the height of a child by 
 means of an eight-foot rule, we do not expect the chilf? 
 to measure up to the eight feet.
 
 28 
 
 Scientific Measurement 
 
 The achievement of a class is measured by calculating 
 the median number of problems which were solved cor- 
 rectly. By the median number is meant that number 
 which marks the point at which there are just as many 
 pupils who solve a greater number correctly as there are 
 those who solve a less number correctly. In order to 
 measure the median point of achievement of the class, 
 it is necessary to make a distribution table, showing the 
 number of pupils who were unable to solve a single prob- 
 lem correctly, the number who solved one, two, three, 
 etc., up to the final number. Take the following as an 
 example : 
 
 Number of Times a Given Number of Addition Problems Was 
 
 Solved Correctly 
 
 No. of pupils 
 
 012 
 
 1102 
 
 10 
 
 11 
 
 11 
 
 12 
 
 15 
 4 
 
 16 
 
 1 
 
 18 
 
 That is, one pupil failed to solve a single problem. With 
 that exception there were no children who did not solve 
 at least one problem correctly. Two children solved 
 two problems correctly, three children solved three 
 problems correctly, four solved five correctly, and so on. 
 Since there are, let us say, 52 individuals in a given 
 class, " the median point evidently falls between the 
 achievements of the 26th and 27th pupils. Let us begin 
 with the individual who was unable to solve a single 
 problem correctly and count the two individuals who 
 solved two problems, the three who solved three prob- 
 lems, and so on until we come to the step that includes 
 the 26th individual. Now if we are to indicate the exact 
 point in the achievement of the pupils where there are 
 just as many pupils who solve a greater number of prob- 
 lems as there are those who solve a less number, it is 
 necessary to count 5 of the 6 individuals who solved 10
 
 Arithmetic Scales 
 
 29 
 
 problems correctly. Thus, on the assumption that the 
 individuals are distributed over any step at equal dis- 
 tances from one another, the median point is f of the 
 distance through this step. Hence, the median achieve- 
 ment of this class, i.e. the median number of problems 
 solved, is 10.8 problems correctly solved." 
 
 TENTATIVE STANDARDS OF ACHIEVEMENT 
 
 The following standards of achievement have been 
 determined on the basis of tests made on several thousand 
 children from the second to the eighth grades of various 
 school systems. It is possible that with further experi- 
 mentation they may need to be slightly altered. 
 
 Tentative Standard of Achievement for Series B 
 
 Grade 
 
 Addition 
 
 Subtraction 
 
 Multiplica- 
 tion 
 
 Division 
 
 II 
 III 
 
 IV 
 
 V 
 
 VI 
 
 VII 
 
 VIII 
 
 
 
 
 4.5 
 
 9 
 11 
 14 
 16 
 18 
 18.5 
 
 3 
 
 6 
 
 8 
 10 
 12 
 13 
 14.5 
 
 3.5 
 
 7 
 11 
 15 
 17 
 18 
 
 3 
 5 
 
 7 
 10 
 13 
 14 
 
 The standards are based upon the total number of 
 problems that were correctly solved in each grade. Thus 
 in the second grade in addition, the median achievement 
 was 4.5 problems, in the third grade, 9.0 problems, etc. 
 
 All that is necessary, therefore, to test a class is to 
 procure the standard blanks, follow the detailed instruc- 
 tions in administration and scoring, and then determine 
 the median score by the method shown. This median 
 score can then be compared with the tentative scores 
 given by the author. It should be noted of course that 
 these tentative scores would cease to have significance if, 
 previous to the test, the children had been drilled on ex- 
 amples framed with the particular scale problems in mind.
 
 30 Scientific Measurement 
 
 EXERCISES 
 
 1. How does the general character of the work of your class, as 
 revealed by the administration of Test No. 7, Series A, compare with 
 that of other classes of the same grade in your building or city? 
 
 2. How does the work of your pupils in the various sub-branches, 
 as revealed by the tests, compare with the standard scores for your 
 grade? How does it compare with the work in these sub-branches 
 in other classes and schools where you may be able to test? 
 
 3. Suppose the frequent administration of the tests failed to re- 
 veal a reasonable amount of improvement in the various sub-branches, 
 what would this seem to indicate? 
 
 4. Could the tests be utilized to remedy this condition? 
 
 5. What two important facts in regard to the ability of the pupils 
 in your class have the tests revealed? 
 
 6. Suppose the tests showed the ability of a pupil to differ greatly 
 in the various sub-branches, what action should the teacher take in 
 regard to it? 
 
 7. For what purposes may the Woody Scale be used to greater 
 advantage than the Courtis Scale ? 
 
 8. In your experience with the tests, have they tended to show 
 any relation between ability in one branch and ability in another? 
 
 9. What precautions should be taken in administering the tests? 
 10. In what ways should the continued use of these tests increase 
 
 the efficiency of a teacher of arithmetic?
 
 CHAPTER III 
 
 HANDWRITING SCALES 
 
 I. THORNDIKE SCALE 
 II. AYRES SCALE 
 HI. COURTIS TESTS 
 
 Probably there is no subject about which opinions of 
 efficiency are more vaguely expressed than the subject of 
 handwriting. Such terms as "good," "fair," "poor," 
 etc., merely express the individual teacher's judgment as 
 determined by certain factors, such as legibility, grace, 
 character, etc., or by certain styles, such as vertical or 
 slanting, to which that individual is partial. No two 
 teachers mean the same quality by the use of the same 
 term. Consequently such judgments, because they are 
 not expressed in terms of a universal standard which 
 conveys the same meaning to everybody, are of little 
 value when comparisons are necessary. Within recent 
 years attempts have been made to eliminate this unscien- 
 tific type of judgment, which is the natural result of the 
 lack of a standard, by the construction of a scale for 
 measuring the quality of handwriting. Thorndike and 
 Ayres have each devised such a scale or standard, while 
 Courtis has outlined a method by which it is possible to 
 obtain samples of children's handwriting, made under 
 uniform conditions. Each of these methods will be de- 
 scribed briefly in turn. 
 
 31
 
 
 p 
 
 w 
 
 o 
 
 fa 
 
 fa 
 
 o 
 
 CO 
 
 fa 
 P 
 
 5 
 
 o 
 
 e 
 
 w 
 
 i
 
 CO 
 
 33
 
 -6 
 
 3 
 
 f\i 
 
 l*> 
 
 T 
 
 $o 
 
 f 
 
 A 
 
 ^ 
 
 0° 
 
 '{i 
 
 00 
 
 34
 
 <f 
 
 <s> 
 
 35
 
 £ .: 
 
 36
 
 37
 
 d 
 
 39
 
 ^ 
 
 v"N 
 
 40
 
 41
 
 42
 
 0> 
 
 T3 
 o 
 
 H 
 
 >» 
 
 to 
 <U 
 
 +-> 
 
 S-. 
 
 O 
 
 o 
 
 QJ 
 
 J3 
 
 Xi 
 
 0) 
 
 o 
 
 73 
 O 
 
 u 
 
 a, 
 
 t* 
 
 0> 
 h 
 
 co 
 
 •a 
 
 o 
 
 CO 
 
 bO 
 C 
 
 'S 
 bo 
 a> 
 g 
 o 
 
 EH 
 
 43
 
 44 Scientific Measurement 
 
 I. THORNDIKE HANDWRITING SCALES 
 
 Thorndike was the first to construct an objective scale 
 for handwriting. This appeared March, 1910, and was 
 developed as follows. One thousand samples of hand- 
 writing, ranging from the worst to the best to be found 
 in the sixth, seventh, and eighth grades, were given in turn 
 to forty competent judges. Each of these judges was 
 asked to rank these samples according to their "general 
 merit," which was to be based on a combination of grace 
 and legibility, by placing each specimen in one of eleven 
 arbitrary groups in order of increasing merit. Previous 
 experiments had shown that these samples, instead of 
 falling into a thousand different classes, naturally fell into 
 about eleven groups, all the members of a group being of 
 about equal merit. That is, the same thing is true of 
 handwriting as is true of attempting to divide into a 
 thousand classes a thousand people whose height varies 
 from five to six feet. Many would be so nearly of the 
 same height as to make such a classification impracticable, 
 if not impossible. Similarly, exact classification would 
 be impossible in the case of writing, where the distinction 
 between the samples was not pronounced. 
 
 After each judge had placed each sample three or four 
 times in this way in one of these eleven groups, the aver- 
 age result of his rankings was taken as his final grading 
 for each specimen; that is, if a judge ranked a certain 
 specimen of handwriting in class 10 on the first occasion, 
 in class 11 on the second, in class 12 on the third and in 
 class 10 on the fourth, on the whole he placed it some- 
 where between classes 10 and 11, or to be exact, at a 
 point which can be represented by 10.7. Then the re- 
 turns of all the judges were massed and the average of all 
 rankings given to each sample was determined. In this 
 way the place assigned to each specimen by the com- 
 bined opinion of all the judges was fixed. When the 
 averaged judgments were collected (as might be expected
 
 Handwriting Scales 45 
 
 where so many samples were concerned), it was found 
 that some samples were placed in, or approximately in, 
 each of the eleven groups; that is, some samples were 
 graded 1, 2, 3, 4 ... 11, while many samples were given 
 rankings midway between the different groups, indicated 
 by the markings 1.4, 1.6, 2.1, 2.8, etc. 
 
 Now when it is recalled that each one of these groups, 
 in the opinion of the judges, is separated from the others 
 by equal steps of merit, it may readily be seen how a 
 handwriting scale can be obtained, provided only that 
 samples be graded exactly or approximately as falling 
 into groups 1, 2, 3, 4, . . . 11, the handwriting samples in 
 group 2 being as much superior to those in group 1 as 
 those in group 3 are superior to those in group 2, etc. 
 In this way the Thorndike Scale was obtained, a scale 
 whose steps of difference forty competent judges have 
 considered to be equal. Later, this scale was extended to 
 include fifteen classes of handwriting which ranged in 
 quality from handwriting which may barely be called 
 such to that suitable for decorative purposes. 
 
 This scale with its various classes of handwriting has 
 from one to three different styles of writing in each group. 
 Undoubtedly it would be far more satisfactory if each 
 class contained samples of all the various types of writing 
 which are found in the scjidol. This defect, however, 
 can easily be remedied when a larger number and greater 
 variety of samples become available. Furthermore, it is 
 to be regretted that this scale, which measures about 
 twenty-two by twenty-four inches, is not issued in more 
 convenient form. 
 
 In spite of these slight defects, which time will remedy, 
 the scale is certainly far superior to the judgment of any 
 one individual. The method of using it is very simple. 
 A sample of handwriting is measured by placing it along- 
 side the scale and estimating to which one of the fifteen 
 groups, as represented by the fifteen samples, it belongs.
 
 46 Scientific Measurement 
 
 If it is thought to lie between two groups, a fraction may 
 be added or subtracted according to whether it is judged 
 better or worse than the sample on the scale to which 
 it most nearly corresponds. Thus, if it falls between 
 classes 12 and 13, it might be graded at any point in 
 between, such as 12.4 or 12.8. For especially accurate 
 work, it is well to have several individuals rank the 
 samples of handwriting and then take the average of 
 their rankings as the final measurement. Care should be 
 taken to decide a specimen's grade not because of its like- 
 ness in style to some sample in the scale, but because of its 
 likeness in quality. 
 
 After the person grading has become familiar with the 
 scale, comparisons will be facilitated if the scale is folded 
 so that the samples form the pages of a book. Then the 
 judge should pass rapidly from the lowest to the highest 
 sample, rating the specimen by his impression as a whole, 
 inasmuch as such an impression is the resultant effect of 
 all the qualities possessed by the writing. Long, pains- 
 taking comparisons prevent accuracy instead of securing 
 it. When it is necessary to compare a specimen with 
 samples unlike it in slant and character, placing it some- 
 where between two groups will often solve the difficulty. 
 
 H. AYRES HANDWRITING SCALE 1 
 
 The Thorndike Scale is based on general merit of hand- 
 writing. The Ayres Scale, on the other hand, is based on 
 legibility; thus there is a substitution of function, instead 
 of appearance, as a criterion. Ayres takes this standard 
 for two reasons. In the first place, the purpose of writing 
 is to be read ; hence " readability," or legibility, is the 
 prime requisite. In the second place, it is exceedingly 
 easy to measure the legibility of any sample of hand- 
 writing by determining the time it takes to read it. In this 
 
 1 For reproduction of Ayres Scale, see pages 50 to 57.
 
 Handwriting Scales 47 
 
 way an exact evaluation of the relative legibility of any 
 specimens may be obtained in terms of a unit of time. 
 The criterion of general merit, though based on the opinion 
 of competent judges, does not allow of this accuracy. 
 
 The method by which this scale was produced differs 
 radically from that used by Thorndike. Previous experi- 
 ments had shown that the best way to find out the rela- 
 tive legibility of different samples of handwriting was to 
 find out the rate, in words per minute, at which each 
 sample could be read. In order to represent a random 
 selection and not the writing of any particular city or 
 section, 1578 samples were secured of the handwriting of 
 children in the upper elementary grades of 40 school 
 systems in 38 different states. These samples did not 
 consist of words so arranged as to convey a meaning, but 
 were composed of words thrown out of context. The 
 object of this was to make it necessary for the reader to 
 decipher each word separately, and to make it impossible 
 for him to memorize. Through the cooperation of super- 
 intendents and teachers, samples from either the best or 
 the worst class in any city were avoided, and it was so 
 arranged that the pupils made no effort to write with 
 exceptional care or rapidity. 
 
 These 1578 samples were then turned over to ten com- 
 petent paid assistants, who in turn read each sample and 
 by means of a stop watch recorded the exact time it took 
 to read it. After each sample had been read by the ten 
 readers, the average time taken to read it was computed. 
 Then the rate in words per minute at which the reading 
 had been accomplished, was found by dividing the average 
 time it took to read a given sample by the number of 
 words in it. This process was repeated for each one of 
 the 1578 samples. After it had been determined to what 
 extent the readers had increased their reading speed 
 through practice, the first 75 papers were reread and new 
 times recorded to correct this error.
 
 48 Scientific Measurement 
 
 The next step was the classification of these samples. 
 After various attempts at this, five classes —vertical, 
 medium slant, extreme slant, backhand, and mixed — 
 were finally carefully defined on the basis of the arbitrary 
 judgment of a number of competent judges. Then each 
 of the samples was classified on the basis of the slant of 
 its letters and assigned to one of these five classes. Be- 
 cause of the limited number of backhand and mixed 
 samples of handwriting, these were left out of the final 
 scale. 
 
 The scale itself was then constructed in the following 
 manner. All of the samples, which had been so marked 
 as to indicate both the rate at which each one had been 
 read and the class or style of writing — vertical or slant, 
 etc. — to which it belonged, were arranged in one long 
 series beginning with the sample having the lowest time 
 rating and extending to the one having the highest. As 
 might be expected, there were many samples of medium 
 grade — that is, that were read at a medium rate ; only 
 a few that were very good — read at a rapid rate; and 
 only a few that were poor — read at a very slow rate. 
 Then, beginning at the poorest sample — that which took 
 the longest time to read — a count was made just halfway 
 through the samples. The specimen thus obtained was 
 the central point, below which one half of the samples 
 were read more slowly, and above which, one half were 
 read more rapidly. This sample had been marked 175.7 
 indicating that it was read at the rate of 175.7 words per 
 minute. Because of its central position, considering the 
 entire series as 100, this sample was called 50. In a simi- 
 lar manner samples were picked out one tenth, two tenths, 
 three tenths, four tenths, six tenths, seven tenths, eight 
 tenths, and nine tenths of the way through the series, 
 and these were designated 10, 20, 30, 40, 60, 70, 80 and 
 90, respectively. These values were chosen because 
 teachers are familiar with them in grading.
 
 Handwriting Scales 49 
 
 The rate of reading marked on these samples was 
 found to be 130.2, 149.4, 163.5, 175.7, 186.1, 195.8, 202.9, 
 and 209.6 words per minute, respectively. Thus it was 
 seen that this scale does not proceed by equal steps as far 
 as the time consumed in reading is concerned. Instead, 
 the gain in time rate became progressively smaller as one 
 moved from the worst to the best sample. How reasonable 
 this is, may easily be seen. A very poor handwriting takes 
 a long time to decipher. One which is just slightly better 
 may be read almost twice as fast. A still better one may 
 be read somewhat faster but not twice as fast as the pre- 
 vious one, and so on, the gain in the rate growing smaller 
 and smaller as the handwriting improves. Thus, as far 
 as readability is concerned, the difference in the time it 
 takes to read a sample marked 30 and one marked 40 is 
 greater than the difference in the time it takes to read 
 one marked 60 and one marked 70. So, what is actually 
 meant when it is said that the steps of this scale are equal, 
 is that each one of them has been so chosen that it is as 
 much better than the one before, as that one is better 
 than the preceding one. Qualities 60 and 40 are respec- 
 tively equally distant above and below quality 50 ; that 
 is, there is the same proportion of samples between 50 and 
 60 as between 40 and 50, and so on down the scale. 
 
 The scale itself is on a sheet of paper measuring nine by 
 thirty-six inches. It contains eight groups (the lowest 
 and the highest groups being omitted in the final scale), 
 each group including three types of handwriting, the 
 vertical, the slant, and the extreme slant. Ayres' studies 
 have shown that 95% of the common writings of school 
 children are included in these three styles. To facilitate 
 comparison, both the paper and the ink used in the scale 
 are of the color used in the public schools. The scale is 
 used in exactly the same way as the Thorndike Scale. 
 
 The following scales are reproduced by the courtesy of Dr. 
 Leonard P. Ay res.
 
 20 
 
 
 a, 
 
 a 
 
 
 
 50
 
 30 
 
 
 <^/^*^ %&i?^ 
 
 <^^^^' ?^ «^ *s>**^ c*rrt^£L~y* \ 
 
 £Z£? ^ <^cz*??t, tzsyr*u^<*t, 
 
 51
 
 40 
 
 ZeAAATLs Iv^oJt Aa/O^ star s\s4s<O0r*^*' f 
 
 Jiad a*i^ ^<7^^ cud/ 
 
 52
 
 50 
 
 JUA^M sto j-vc^tUj^ CUYrM&L <&*kst 
 
 ]U^Mhcft itfZO^^'&W' 'JltM^nCfi 
 
 
 53
 
 60 
 
 *&c* <tf '^6xel^ ^A*~
 
 70 
 
 QsOm, -fayiX srwr ^DuisCA^ off w 
 
 55
 
 80 1 
 
 <rP /wCc^WNJ2- VvA^ ,A~vOwt> AATYUxX Cur* 
 
 Ascr&sys \^n~<t*4/ jZslp&s, ^esirtSu o- 
 
 
 G^y~tsc£ &sO<s~e 
 
 56
 
 90 
 
 life kadi rurt vum, luyme. Zoy\& ax> 
 
 Off sOW>U^l^^s<*Szi^s£*&^ 
 
 
 *s£<rL£sts / b' sxZst? 
 
 57
 
 58 Scientific Measurement 
 
 m. COURTIS HANDWRITING TESTS 
 
 It is apparent that the efficient use of such objective 
 scales necessitates imposing uniform conditions when 
 obtaining the handwriting samples to be measured. 
 That is, it would be unfair to grade by the same stand- 
 ard a sample of handwriting written at a very rapid rate 
 of speed, and one written at a very slow rate of speed. 
 The time allowed in the writing of the samples should be 
 the same. Courtis has attempted to overcome this diffi- 
 culty by the use of a simple method whereby samples 
 may be obtained of a pupil's handwriting under formal 
 conditions (at a fixed rate of speed), and under casual 
 conditions (at the writer's natural rate of speed) . Samples 
 written under either one of these conditions, providing 
 all the samples were obtained under the same circum- 
 stances, may then be graded by either the Thorndike or 
 the Ayres scale. 
 
 The value of such scales as have been described is very 
 great. In the first place, they represent measures that 
 are the outcome of the thought, labor, and experiment of 
 many persons and as such they are far superior to the 
 casual and oftentimes thoughtless opinion of any one 
 individual. Not only may such scales measure the rela- 
 tive quality of different samples of handwriting or the 
 improvement in handwriting in an individual, but by 
 means of them entire classes, groups, and systems, whether 
 chosen for grade, age, sex, method of teaching, length of 
 practice drills, etc., may be compared ; for the scale con- 
 stitutes the basis of judgment wherever it is used. So it 
 makes no difference whether a sample is called X, A, 
 100, or any arbitrary name, provided it represents in 
 terms of the scale the same quality to every person. By 
 means of these scales teachers may determine the progress 
 and needs of their pupils as well as the efficiency of the 
 methods they are using. Supervision may also be made
 
 Handwriting Scales 59 
 
 more effective because school officials, given the actual 
 results of such tests, are provided with the means for 
 making a scientific analysis of school conditions. How- 
 ever, as Thorndike says, such scales will do their greatest 
 service not as measuring rules, but "by creating in the 
 minds of teachers a mental standard to be used in even 
 the most casual ratings of everyday school life." 
 
 STANDARD SCORES 
 
 It is obvious that if the above scales are used in meas- 
 uring the writing ability of a sufficient number of chil- 
 dren, standard scores of efficiency for each grade may be 
 computed by finding the arithmetical average of the 
 records obtained in each grade. That is, on the basis of 
 the results obtained from the use of the scales, it is pos- 
 sible to say just how fast and how well pupils in any one 
 grade should write. 
 
 Starch secured tentative standard scores of this kind 
 by administering the Thorndike Test at the end of the^ 
 school year to about 4000 pupils in eight cities of .three 
 states, and finding (1) the average rate of speed and (2) the 
 average quality of writing, in each one of the grades. 
 Starch's results were as follows : 
 
 Grades 1 
 
 Speed (letters per minute) 20 
 Quality (Thorndike Scale) 6.5 
 
 According to these scores we find that the average 
 child in the second grade, for instance, writes at a speed 
 of 31 words per minute and possesses a quality of hand- 
 writing indicated by 7.5 on the Thorndike Scale or, by 
 derivation, 26.5 on the Ayres Scale. 
 
 The Thorndike Scale was chosen in preference to the 
 Ayres Scale because the latter does not extend so far at 
 the lower and upper limits as the former; that is, the 
 
 2 
 
 3 
 
 4 
 
 5 
 
 6 
 
 7 
 
 8 
 
 31 
 
 38 
 
 47 
 
 57 
 
 65 
 
 75 
 
 83 
 
 7.5 
 
 8.2 
 
 8.7 
 
 9.3 
 
 9.8 
 
 10.4 
 
 10.9
 
 60 
 
 Scientific Measurement 
 
 limits of the Ayres Scale lie within qualities 7 to 14 of the 
 Thorndike Scale. However, equivalent values for quali- 
 ties in the Ayres Scale have been derived ; that is, one 
 step on the Thorndike Scale has been found to be equal 
 to 8.9 points on the Ayres Scale. So the standard scores 
 may be expressed in units of either scale. As a result of 
 this derivation we have the following equivalents : 
 
 Thorndike Scale 
 
 
 
 Ayres Scale 
 
 Quality 
 
 7 
 
 is equal to 
 
 quality 22 
 
 << 
 
 8 
 
 a n 
 
 it 
 
 31 
 
 <( 
 
 9 
 
 « «< 
 
 tt 
 
 " 40 
 
 <« 
 
 10 
 
 tt tt 
 
 It 
 
 " 49 
 
 <« 
 
 11 
 
 it it 
 
 tt 
 
 58 
 
 tt 
 
 12 
 
 it ii 
 
 tt 
 
 67 
 
 n 
 
 13 
 
 ii it 
 
 tt 
 
 76 
 
 << 
 
 14 
 
 it tt 
 
 tt 
 
 " 85 
 
 In obtaining these standards it was apparent (1) that 
 writing in the various grades differed widely in regard to 
 both speed and quality ; and (2) that the abilities of the 
 pupils in one grade in many instances overlapped the abil- 
 ities of the pupils in the adjacent grades. For instance, the 
 handwriting of pupils in the first grades in three schools 
 in city A was found to range all the way from quality 4 
 to quality 10.5, and that of pupils in the eighth grade, 
 from quality 5 to quality 15. So great was the over- 
 lapping that the averages of the various grades differed 
 from each other by only small amounts. 
 
 To recapitulate, the first thing to be done by a teacher 
 who is desirous of measuring the handwriting of pupils, 
 is to see that the samples to be measured are all obtained 
 under the same conditions. This may be done by having 
 the pupils write at their natural rate of speed a selection, 
 or part of a selection, with which they are all familiar. 
 Care should be taken to see that they all start writing at 
 the same time and stop at the same time, say at the end 
 of two minutes.
 
 Handwriting Scales 61 
 
 The speed of writing for each pupil may be easily deter- 
 mined, in terms of letters per minute, by dividing the total 
 number of letters written by two. The average speed of 
 writing for the class may then be computed and com- 
 pared with the standard score for that grade as given by 
 Starch. 
 
 The samples may then be measured for their quality 
 by either the Thorndike or the Ayres Scale. The former 
 may be secured by sending to Teachers College, Columbia 
 University, New York ; and the latter, by sending to the 
 Russell Sage Foundation, Department of Education, New 
 York. 
 
 In determining the quality of any given specimen of 
 handwriting, all that is necessary is to slide the specimen 
 along the scale beginning with the poorest sample until 
 a writing of corresponding quality is found. If this 
 writing is marked, say 11 in the Thorndike Scale, then 
 that is the value of the handwriting measured. If the 
 quality of the specimen seems to be better than quality 
 11, but not so good as quality 12, then it should be given 
 a value somewhere between 11 and 12, say 11.2 or 11.8, 
 according to whether it is judged better or worse than the 
 sample on the scale to which it most closely corresponds. 
 
 It is advisable for the teacher to have one of the scales 
 on the wall, in order that it may be utilized by the pupils 
 themselves. 
 
 It is obvious that in comparing the handwriting of any 
 two classes of the same grade the following conditions 
 must have been fulfilled : (1) All of the children must 
 have written the same selection for their samples ; (2) all 
 must have had the same degree of familiarity with the 
 selection ; (3) all must have been allowed the same time 
 in which to write ; (4) all the results must be expressed in 
 terms of the same scale. Such uniform conditions may 
 be realized within any given school system through a 
 set of specific instructions sent out by the supervisor of
 
 62 Scientific Measurement 
 
 writing or the superintendent of instruction some days 
 before the measurements are to be made. 
 
 EXERCISES 
 
 1. Take thirty specimens of writing distributed through the 
 grades and have these marked by five teachers according to their 
 usual percentage methods. How do the results agree? 
 
 2. Repeat the above experiment with the exception that the 
 grading is done by means of the Thorndike Scale. How do the esti- 
 mates of the five teachers compare now? 
 
 3. Suppose a teacher found as a result of using the scale that the 
 pupils were above the average (standard score) for that grade in 
 speed of writing but below the average in quality, or vice versa, what 
 should be done? 
 
 4. Does there seem to be any relation between speed and quality 
 of handwriting? 
 
 5. Is there any practical lesson to be learned from the fact that, 
 although a group of children may be made to write in the same way 
 up to 14 years of age, at 18 each has his own particular style? 
 
 6. What would be the result of getting the children to grade 
 their own handwriting and to compare their results? 
 
 7. Suppose a teacher found a great difference in the quality of 
 the handwriting of the pupils in a given grade, some pupils writing 
 far above the average for that grade and others far below it, what 
 should be done? 
 
 8. What is the best thing to do if a teacher finds it necessary to 
 compare a given specimen of handwriting with samples in the scale 
 which are unlike it in slant or character, or both? 
 
 9. How may the scale be used to test the efficiency of any method 
 of teaching handwriting? 
 
 10. What factors should a teacher take into consideration when 
 setting a standard of handwriting for a given class?
 
 CHAPTER IV 
 READING SCALES 
 
 I. THORNDIKE AND GRAY SCALES 
 H. STARCH SCALE 
 HI. COURTIS SCALE 
 
 Since it is through reading that a large part of our in- 
 formation is obtained, an objective means of measuring 
 efficiency in that subject is of great importance. Several 
 attempts have been made to fill this need. In 1914 
 Thorndike and Gray published tentative scales for measur- 
 ing school achievement in reading, and later both Starch 
 and Courtis published tests for the same purpose. 
 
 I. THORNDDIE AND GRAY SCALES 
 
 In Thorndike and Gray's Scales attempts are made to 
 measure the following factors : " (1) silent reading so far 
 as it concerns (a) the understanding of words singly and 
 (6) the understanding of sentences and paragraphs ; and 
 (2) simple, oral reading of matter-of-fact passages." 
 Each of these three scales will be described in turn. 
 
 (i) Scale A for Visual Vocabulary 
 
 The scale designed to measure a pupil's knowledge of 
 the meaning of single words is called "Scale A or the 
 Scale for Extent or Range of Visual Vocabulary." It is 
 printed on a single sheet of paper and consists of nine 
 lines of words. These lines are numbered 4, 5, 6, 7, 8, 
 9, 10, 10.5 and 11 respectively. All the words on the 
 same line are about equally hard to understand, and their 
 difficulty increases gradually from line to line. The first 
 
 63
 
 64 Scientific Measurement 
 
 line is marked 4 because the difference between a child 
 who can read the first line and one who can read nothing 
 at all is about four times as great — measured in years 
 of work — as the difference between a child who can read 
 line 4 and one who can read line 5, or as the difference 
 between pupils who can read any other two successive 
 lines. The seventh line is marked 10.5 because the words 
 on it are a little too hard to stand half-way between lines 
 9 and 11. There are only three words on line 11 because 
 no others of precisely the same difficulty could be found. 
 
 Thorndike Reading Scale A 
 Visual Vocabulary 
 
 Write your name here 
 
 Look at each word and write the letter F under 
 
 every word that means a flower. 
 Then look at each word again and write the letter 
 
 A under every word that means an animal. 
 Then look at each word again and write the letter 
 
 N under every word that means a boy's name. 
 Then look at each word again and write the letter 
 
 G under each word that means a game. 
 Then look at each word again and write the letter 
 
 B under every word that means a book. 
 Then look at each word again and write the letter 
 
 T under every word like now or then that means 
 
 something to do with time. 
 Then look at each word again and write the word 
 
 GOOD under every word that means something 
 
 good to be or do.
 
 Reading Scales 65 
 
 Then look at each word again and write the word 
 BAD under every word that means something 
 bad to be or do. 
 
 4. camel, samuel, kind, lily, cruel 
 
 5. cowardly, dominoes, kangaroo, pansy, tennis 
 
 6. during, generous, later, modest, rhinoceros 
 
 7. claude, courteous, isaiah, merciful, reasonable 
 
 8. chrysanthemum, considerate, lynx, prevari- 
 cate, reuben 
 
 9. ezra, ichabod, ledger, parchesi, preceding 
 
 10. crocus, dahlia, jonquil, opossum, poltroon 
 
 10.5 begonia, equitable, pretentious, renegade, 
 reprobate 
 
 11. armadillo, iguana, philanthropic
 
 66 Scientific Measurement 
 
 The child's score or measure is determined by finding 
 the hardest, or the highest-numbered, line that he marks 
 with not more than a single error — all omissions being 
 regarded as errors. The number of this line is taken as 
 the child's score. For example, five children gave results 
 as follows : 
 
 Number of Omissions and Errors in Each Line in the Case 
 of Five Pupils, C, J, N, R, and W 
 
 Line 4 5 6 7 8 9 10 10.5 11 
 
 Pupil C 000013 3 4 3 
 
 "J 000011 1 3 2 
 
 "N 000124 4 3 3 
 
 "R 000012 3 5 3 
 
 «W 000000 1 1 
 
 Thus we may say that C has ability 8 ; J has ability 
 10; N has ability 7; R has ability 8; W has ability 
 10.5 or 11 or better. 
 
 To measure the ability of a class as a whole, we simply 
 take an average by adding together all the errors and 
 omissions on each line and dividing by the number of 
 children. In a rough estimate the class gets credit for 
 the highest-numbered line that shows an average error 
 of 1 or less, and the figure thus obtained, or the result as 
 a whole, can be used for comparison with the achievement 
 of other classes. 
 
 Considering the five children above mentioned as a 
 class, we have, as the average number of errors and 
 omissions on each line, the following : 
 
 Line 456 7 8 9 10 10.5 11 
 
 Errors (including 
 omissions) . . .2 1.0 2.0 2.4 3.2 (2.2 
 
 for the three-word line 11 or 3.7 for 
 a five-word line of equal difficulty)
 
 Reading Scales 67 
 
 Since the highest-numbered line that this class marked 
 with not more than an average of one error or omission 
 per child is line number 8, 8 may be considered the score 
 or measure of this class. 
 
 The choice of four out of five correct as a standard 
 could be replaced by all correct (100%) or three correct 
 (60%), but for statistical reasons 80% is the best criterion. 
 
 The measures procured by this method are not only 
 objective, but they have a definite meaning. For in- 
 stance, to say that an individual or class possesses ability 
 6 in reading, means that he or they possess the ability to 
 mark correctly at least four out of five (80%) of the words 
 in line 6 on the scale. Furthermore, the difference in 
 difficulty between lines 4 and 5 is approximately equal 
 to the difference in difficulty between lines 5 and 6, and 
 so on. Lastly, since the difference in difficulty between 
 lines 8 and 4 is probably about equal to that between 4 
 and 0, the attainment of a class scored 8 may be said to 
 be about twice as great as that of one scored 4. 
 
 In the case of an individual, it is always possible to 
 state in just what line are recorded errors and omissions 
 totaling 20% or less (80% or more correct) ; that is, 
 the individual's " degree of difficulty " can be accurately 
 stated at sight. When it is a matter of an entire class, 
 however, this is not so easy. For instance, a class may 
 have a record of 16% of errors and omissions for line 6, 
 and 25% for line 7. In such a case the " degree of diffi- 
 culty " which would give a percentage of 20 may be 
 obtained by consulting the tables and following the 
 directions in the original paper. In short, when the 
 percentage of errors or omissions for a given line, say 
 6, is known, it is possible to estimate just the " degree 
 of difficulty," say 6.7, which would give a percentage 
 of exactly 20 (80% correct) for the class in question. 
 
 The time required to measure a class of forty pupils, 
 record the results, and estimate the " degree of difficulty "
 
 68 Scientific Measurement 
 
 that would give a percentage of 20 errors and omissions, 
 varies from two to five hours. Thorndike believes it 
 would be well for every school, from the fourth to the 
 eighth grades, to make such measurements at the begin- 
 ning and at the end of the school year. 
 
 This scale is not without defects and limitations, some 
 of which Thorndike is at present working to overcome. 
 (1) A scale whose steps or lines contain 10 or 20 words, 
 instead of 5, will obviously give data for a more precise 
 estimate of the ability of a class. (2) When applied to a 
 single pupil the scale is not so precise as when applied 
 to a class, for a child who happened to be interested in 
 flowers and animals would have a decided advantage 
 over one who was not so interested in them. (3) A pupil's 
 score cannot always be exactly stated; for, if a child 
 misses 2 words in line 8, no words in line 9 and 3 words 
 in line 10, shall his ability be classed as 7 or 9? How- 
 ever, a reasonably rough estimate of his ability may be 
 gained by consulting his score in the other lines. In a 
 class, the chance familiarity of a pupil with certain 
 words will be counterbalanced by the chance unfamiliar- 
 ity of some other pupil. (4) The fact that words ex- 
 pressing relations, such as pronouns, prepositions, and 
 conjunctions, are omitted in this scale, seems a serious 
 limitation until it is considered that the chief importance 
 of these words is in sentence comprehension, and that 
 the scale for that purpose, which will be described later, 
 tests knowledge of them rather thoroughly. (5) Not all 
 the words on a given line are of absolutely equal difficulty, 
 but the differences in the degree of difficulty are not of 
 enough importance to constitute a defect. (6) It must 
 be admitted that the differences between successive 
 lines are not exactly equal. In fact, even " their ap- 
 proximate equality depends upon the approximate truth 
 of certain hypotheses about the distribution of word- 
 knowledge in children of the same grade and about the
 
 Reading Scales 69 
 
 comparative variability of the children in Grades IV, V, 
 VI, VII, and VIII in respect to word-knowledge." 
 (7) Lastly, it must be remembered that this scale does 
 not measure the meaning of the printed words, save as 
 required in the directions on the scale. 
 
 In spite of the difficulties which any such scale pre- 
 sents, it may be used for practical purposes, at its face 
 value. It is capable of revealing large individual differ- 
 ences within a class and of measuring them roughly, if, 
 as Thorndike says, it is interpreted with common sense. 
 Moreover, as a measure of " the ability to understand 
 printed words unconfused with the ability to express 
 one's self orally or in writing," it is superior to any form 
 of definition test. By extending it to include more 
 difficult words it may be used to measure achievement 
 and improvement from the third grade through college. 
 Indeed, with slight modification it can be used to measure 
 extent of vocabulary in any foreign language, and in 
 fact, such scales for French, German, and Latin are being 
 planned. 
 
 Scale A is designed for use in Grades IV to VIII in- 
 clusive in the elementary schools and to some extent 
 in the high schools. To be sure that the general nature 
 of the scale is understood, a short, simple, preliminary 
 test, similar in character to Scale A, should be given. 
 A pupil who has made less than five errors and omissions 
 in the first two lines taken together in the preliminary 
 test may be assumed to understand the general idea of 
 the scale. A pupil in the third grade or above who makes 
 more than ten errors and omissions in the first two lines 
 taken together, may be assumed not to understand what 
 is required of him. In the fourth grade one half an 
 hour should be allowed for the test, in the fifth and sixth 
 grades twenty-five minutes, and in the seventh and 
 eighth grades, twenty minutes. Although a time record 
 is not used in the measurement of the vocabulary itself,
 
 70 Scientific Measurement 
 
 Thorndike believes that it should be kept, without the 
 pupil's knowledge, since it will prove instructive and 
 requires little labor. A little experience will soon teach 
 the scorer what lines he need score for a given class. 
 For instance, in the eighth grade lines 4, 5, and 6 may 
 almost always be neglected, while in the fourth grade 
 lines 10, 10.5 and 11 may safely be disregarded. 
 
 The words of Scale A were chosen from a much larger 
 number which were tried upon about 2500 pupils in the 
 fourth, fifth, sixth, seventh and eighth grades in five 
 different schools. Words were considered to be of ap- 
 proximately equal difficulty if approximately equal per- 
 centages of pupils in the fourth, fifth, sixth, seventh and 
 eighth grades, respectively, marked them correctly in 
 these tests. For instance, the words finally selected 
 for row 4 in the scale — camel, samuel, kind, lily, and 
 cruel — were marked correctly by approximately the same 
 percentage of pupils in all the fifth grade classes, and sim- 
 ilarly in the other grades ; that is, about 100% got each of 
 them right in the eighth grade, about 99% in the seventh 
 grade, 98% in the sixth grade and 96% in the fifth grade. 
 At present it is planned to improve the scale so that 
 each row will include ten words instead of five. Some 
 of the words added will be similar to those already in 
 the scale, such as boys' names, while others will be words 
 of equal difficulty obtained by administering new tests. 
 An attempt will also be made to find words of difficulty 
 11.5, 12 and 12.5. With the material which he has 
 already collected, Thorndike expects to enlarge Scale A 
 by adding words of difficulty 4.5, 5.5, 6.5, 7.5, 8.5 and 9.5. 
 In this way the exactitude of measurement of extent or 
 range of visual vocabulary will be greatly increased.
 
 Reading Scales 71 
 
 (2) Scale Alpha. For Measuring the Understanding 
 
 of Sentences 
 
 Thorndike's second scale, Scale Alpha, is an attempt 
 to measure the ability of a child to read understandingly, 
 that is, to understand the meaning of sentences and 
 paragraphs. The value of such a scale is obvious when 
 it is realized that competent judges would rate this 
 ability "at from 60% to 90% of the total result to be 
 sought by the elementary school in the teaching of 
 reading." 
 
 In constructing this scale, preliminary experimentation 
 was conducted along two separate lines ; namely, (1) meas- 
 urement by the passage-question method and (2) measure- 
 ment by responses in marking letters, numbers, and the 
 like. The work in both of these lines was so successful 
 in measuring the pupil's ability to read understandingly, 
 that the two types of measurement were employed in the 
 final scale, which consists of four " sets " or steps, each 
 one of which contains from one to five questions. This 
 scale is reproduced below. 
 
 SET a or 4 
 
 Read this and then write the answers. Read it again 
 as often as you need to. 
 
 John had two brothers who were both tall. 
 Their names were Will and Fred. John's sister, 
 who was short, was named Mary. John liked 
 Fred better than either of the others. All of these 
 children except Will had red hair. He had brown 
 hair. 
 
 1. Was John's sister tall or short ? 
 
 2. How many brothers had John ? 
 
 3. What was his sister's name ?
 
 72 Scientific Measurement 
 
 SET b or 6 
 
 Read this and then write the answers. Read it again 
 as often as you need to. 
 
 Long after the sun had set, Tom was still wait- 
 ing for Jim and Dick to come. " If they do not 
 come before nine o'clock," he said to himself, " I 
 will go on to Boston alone." At half past eight 
 they came bringing two other boys with them. 
 Tom was very glad to see them and gave each of 
 them one of the apples he had kept. They ate 
 these and he ate one too. Then all went on down 
 the road. 
 
 1. When did Jim and Dick come ? 
 
 2. What did they do after eating the apples? 
 
 3. Who else came besides Jim and Dick ? 
 
 4. How long did Tom say he would wait for them ? 
 
 5. What happened after the boys ate the apples ?
 
 Reading Scales 73 
 
 SET c or 8 
 
 Read this and then write the answers. Read it again 
 as often as you need to. 
 
 It may seem at first thought that every boy and 
 girl who goes to school ought to do all the work 
 that the teacher wishes done. But sometimes 
 other duties prevent even the best boy or girl 
 from doing so. If a boy's or girl's father died 
 and he had to work afternoons and evenings to 
 earn money to help his mother, such might be 
 the case. A good girl might let her lessons go 
 undone in order to help her mother by taking care 
 of the baby. 
 
 1. What are some conditions that might make 
 even the best boy leave school work unfinished ? 
 
 2. What might a boy do in the evenings to help 
 his family ? 
 
 3. How could a girl be of use to her mother? . . . . 
 
 4. Look at these words : idle, tribe, inch, it, ice, 
 ivy, tide, true, tip, top, tit, tat, toe. 
 
 Cross out every one of them that has an i and 
 has not any t (T) in it.
 
 74 Scientific Measurement 
 
 SET d or 10 
 
 Read this and then write the answers. Read it again 
 as often as you need to. 
 
 It may seem at first thought that every boy and 
 girl who goes to school ought to do all the work 
 that the teacher wishes done. But sometimes 
 other duties prevent even the best boy or girl 
 from doing so. If a boy's or girl's father died and 
 he had to work afternoons and evenings to earn 
 money to help his mother, such might be the case. 
 A good girl might let her lessons go undone in 
 order to help her mother by taking care of the 
 baby. 
 
 1. What is it that might seem at first thought to 
 be true, but really is false ? 
 
 2. What might be the effect of his father's death 
 upon the way a boy spent his time ? 
 
 3. Who is mentioned in the paragraph as the per- 
 son who desires to have all lessons completely 
 done ? 
 
 4. In these two lines draw a line under every 5 
 that comes just after a 2, unless the 2 comes 
 just after a 9. If that is the case, draw a line 
 under the next figure after the 5 : 
 536254174257654925386125 
 473523925847925612574856 
 
 The foregoing scales and tables in this section are reproduced 
 by the courtesy of Dr. E. L. Thorndike.
 
 Reading Scales 75 
 
 In the first two sets or steps — "a or 4 " and "b or 
 6 " — of the scale just given t the first type of measure- 
 ment or the passage-question method, is used. Here 
 the ability to understand a sentence or paragraph is 
 measured by the correctness of verbal responses to cer- 
 tain questions asked regarding it. In the last two sets 
 or steps — " c or 8 " and " d or 10 " — ability to under- 
 stand a sentence or short paragraph is measured by re- 
 sponses which are not entirely verbal in character, such 
 as marking letters and numbers. Each one of the four 
 " sets " or steps is more difficult than the preceding 
 one. 
 
 As in the case of Scale A, a preliminary test should 
 be given the pupils before administering Scale Alpha, to 
 find out if they understand instructions. Scale Alpha is 
 available for Grades III to VIII. Twenty to thirty 
 minutes should be allowed for administering it, and 
 scoring is done as in Scale A. In marking the responses 
 " the general intent should be to require an answer that 
 proves that the pupil has understood the passage per- 
 fectly." Because of the small number of steps in the 
 scale, the " degree of difficulty " or, what amounts to 
 the same thing, the ability of an individual or class, may 
 be estimated from the percentage of errors and omissions 
 nearest 20%, in a very similar manner to that used in 
 Scale A; but for detailed directions the original paper 
 must be consulted. 
 
 Thorndike points out that the values for the steps of 
 this scale are not at all exact; that is, the difficulty of 
 Set 4, for instance, is not exactly two and one-half times 
 that of a possible Set 1, but he has permitted this estimate 
 to stand, to facilitate the understanding of the scale. In 
 order to obtain a scale of four or more steps, and make 
 sure that all the questions in each step are of approxi- 
 mately equal difficulty and that there is a uniform differ- 
 ence in difficulty between the different steps, it will be
 
 76 Scientific Measurement 
 
 necessary to test over 4000 pupils, obtaining from each 
 from 50 to 60 responses. 
 
 As Thorndike points out, even though Scale Alpha is 
 but provisional, its use will make comparison much fairer 
 and more exact than hours of oral questioning on the 
 part of the most capable supervisor of reading. The scale 
 will eventually be extended and improved by adding 
 other elements equal in difficulty to those now given 
 and by filling in with intermediate steps. 
 
 (3) The Gray Tentative Scale for Measuring Achievement 
 
 in Oral Reading 
 
 This provisional scale for measuring ability to pro- 
 nounce English sentences consists of ten paragraphs of 
 increasing reading difficulty. 
 
 Passage a 
 
 It was time for winter to come. The little birds 
 had all gone far away. They were afraid of the cold. 
 There was no green grass in the fields, and there were 
 no pretty flowers in the gardens. Many of the trees 
 had dropped all their leaves. Cold winter with its 
 snow and ice was coming soon. 
 
 Passage b 
 
 Once there lived a king and queen in a large palace, 
 but the king and queen were not happy. There were 
 no little children in the house or garden. One day 
 they found a poor little boy and girl at their door. 
 They took them into the palace and made them their 
 own. The king and queen were then happy.
 
 Reading Scales 77 
 
 Passage c 
 
 Once I went home from the city for a summer's 
 rest. I took my gun for a stroll in the woods where 
 I had shot many squirrels. I put my gun against a 
 tree and lay down upon the leaves. Soon I was fast 
 asleep, dreaming of a group of merry, laughing children 
 running and playing about me on all sides. 
 
 Passage d 
 
 One of the most interesting birds which ever lived 
 in my bird-room was a blue jay named Jakey. He was 
 full of business from morning till night, scarcely ever 
 still. He had been stolen from a nest long before he 
 could fly, and he was reared in a house, long before he 
 had been given to me as a pet. 
 
 Passage e 
 
 The part of farming enjoyed most by a boy is the 
 making of maple-sugar. It is better than blackberry- 
 ing and almost as good as fishing. One reason he likes 
 this work is that some one else does most of it. It is a 
 sort of work in which he can appear to be very indus- 
 trious, and yet do but little. 
 
 Passage f 
 
 It was one of those wonderful evenings such as are 
 found only in this magnificent region. The sun had 
 sunk behind the mountains, but it was still light. The 
 twilight glow embraced a third of the sky, and against 
 its brilliancy stood the dull white masses of the moun- 
 tains in evident contrast.
 
 78 Scientific Measurement 
 
 Passage g 
 George Washington was in every sense of the word 
 a wise, good and great man. But his temper was 
 naturally irritable and high-toned. Through reflec- 
 tion and resolution he had obtained a firm and habitual 
 ascendancy over it. If, however, it broke loose its 
 bonds, he was most tremendous in his wrath. 
 
 Passage h 
 Responding to the impulse of habit, Josephus spoke 
 and the others listened attentively, but in grim and 
 contemptuous silence. He spoke for a long time, con- 
 tinuously, persistently and ingratiatingly. Finally ex- 
 hausted through lack of nourishment, he hesitated. 
 As always happens in that contingency, he was lost. 
 
 Passage i 
 The hypothesis concerning physical phenomena for- 
 mulated by the early philosophers proved to be in- 
 consistent and, in general, not universally applicable. 
 Before relatively accurate principles could be estab- 
 lished, physicists, mathematicians, and statisticians had 
 to combine forces and work arduously. 
 
 Passage j 
 Read the following sentences correctly: Sophistry 
 is fallacious reasoning. They resuscitated him. Ver- 
 biage is wordiness. Equanimity is evenness of mind. 
 He has a pertinacious, obstinate disposition. There 
 was subtlety and poignancy in his remarks. A hypo- 
 critical or pharisaical nature is usually cynical. 
 
 The scale and table in this section are reproduced by the courtesy 
 of Mr. W. S. Gray.
 
 Reading Scales 
 
 79 
 
 DIRECTIONS FOR ADMINISTERING 
 
 Pupils are required to read these passages in order, 
 stopping at the end of each paragraph. The gross errors, 
 minor errors, omissions, substitutions, and insertions made 
 in each passage, as well as the time needed to read it, 
 are recorded in detail on a duplicate of the scale. When 
 a child makes 4 or more errors and takes 30 seconds or 
 more to read a given paragraph, or when he makes 5 or 
 more errors, however quickly he reads, he may be con- 
 sidered to have failed to read that passage. Although 
 the difference in the degree of difficulty between any two 
 of these passages has not as yet been definitely established, 
 if values must be assigned to the ten paragraphs, Gray 
 suggests that the following figures be used. 
 
 PA88AGE 
 
 Value 
 
 Passage 
 
 Value 
 
 a 
 
 4.5 
 
 f 
 
 9.5 
 
 b 
 
 5 
 
 g 
 
 11 
 
 c 
 
 6 
 
 h 
 
 12 
 
 d 
 
 7 
 
 i 
 
 14 
 
 e 
 
 8 
 
 J 
 
 15 
 
 When finished, this scale will consist of an exactly 
 graded series compiled from many graded series similar 
 to the one just given. Even in this rough approximation 
 to its final form, the scale is much better than any other 
 means at hand for measuring ability in pronouncing Eng- 
 lish sentences.
 
 80 Scientific Measurement 
 
 H. STARCH READING TESTS 
 No. 1 
 
 Once there was a little girl who lived 
 with her mother. 
 
 They were very poor. 
 
 Sometimes they had no supper. 
 
 Then they went to bed hungry. 
 
 One day the little girl went into the 
 woods. 
 
 She wanted sticks for the fire. 
 
 She was so hungry and sad ! 
 
 "Oh, I wish I had some sweet por- 
 ridge ! ' ' she said. 
 
 "I wish I had a pot full for mother 
 and me. 
 
 We could eat it all up." 
 
 Just then she saw an old woman with 
 a little black pot. 
 
 She said, "Little girl, why are you so 
 sad?" 
 
 "I am hungry," said the little girl.
 
 Reading Scales 81 
 
 No. 2 
 
 Betty lived in the South, long, long 
 ago. She was only ten years old, but 
 she liked to help her mother. 
 
 She had learned to do many things. 
 She could knit and sew and spin ; but 
 best of all she liked to cook. 
 
 One day Betty was alone at home 
 because her father and mother and 
 brother had gone to town to see a won- 
 derful sight. 
 
 The great George Washington was 
 visiting the South. He was going from 
 town to town, riding in a great white 
 coach trimmed with shining gold. It 
 had leather curtains, and soft cushions. 
 Four milk-white horses drew it along 
 the road. 
 
 Four horsemen rode ahead of the 
 coach to clear the way and four others 
 rode behind it. They were all dressed 
 in white and gold.
 
 82 Scientific Measurement 
 
 No. 3 
 
 Little Abe hurried home as fast as his feet could 
 carry him. Perhaps if he had worn stockings and 
 shoes like yours, he could have run faster. But, 
 instead, he wore deerskin leggings and clumsy 
 moccasins of bearskin that his mother had made 
 for him. 
 
 Such a funny little figure as he was, hurrying 
 along across the rough fields ! His suit was made 
 of warm homespun cloth. His cap was made of 
 coonskin, and the tail of the coon hung behind 
 him, like a furry tassel. 
 
 But if you could have looked into the honest, 
 twinkling blue eyes of this little lad of long ago, 
 you would have liked him at once. 
 
 In one hand little Abe held something very 
 precious. It was only a book, but little Abe 
 thought more of that book than he would have 
 thought of gold or precious stones. 
 
 You cannot know just what that book meant 
 to little Abe, unless you are very fond of reading. 
 Think how it would be to see no books except 
 two or three old ones that you had read over and 
 over until you knew them by heart !
 
 Reading Scales 83 
 
 No. 4 
 
 The red squirrel usually waked me in the dawn, 
 coursing over the roof and up and down the sides 
 of the house, as if sent out of the woods for this 
 very purpose. 
 
 In the course of the winter I threw out half a 
 bushel of ears of sweet-corn on to the snow crust 
 by my door, and was amused by watching the 
 motions of the various animals which were baited 
 by it. All day long the red squirrels came and 
 went, and afforded me much entertainment by 
 their maneuvers. 
 
 One would approach, at first, warily through 
 the shrub-oaks, running over the snow crusts by 
 fits and starts like a leaf blown by the wind. Now 
 he would go a few paces this way, with wonderful 
 speed, making haste with his "trotters" as if it 
 were a wager ; and now as many paces that way, 
 but never getting on more than half a rod at a 
 time. 
 
 Then suddenly he would pause with a ludicrous 
 expression and a somerset, as if all eyes in the 
 universe were fixed on him. Then, before you 
 could say Jack Robinson, he would be in the top 
 of a young pitch-pine, winding up his clock and 
 talking to all the universe at the same time.
 
 84 Scientific Measurement 
 
 No. 5 
 
 Once upon a time, there lived a very rich man, and a 
 king besides, whose name was Midas; and he had a little 
 daughter, whom nobody but myself ever heard of, and 
 whose name I either never knew, or have entirely forgotten. 
 So, because I love odd names for little girls, I choose to 
 call her Marygold. 
 
 This King Midas was fonder of gold than anything else 
 in the world. He valued his royal crown chiefly because 
 it was composed of that precious metal. If he loved any- 
 thing better, or half so well, it was the one little maiden 
 who played so merrily around her father's footstool. But 
 the more Midas loved his daughter, the more did he desire 
 and seek for wealth. He thought, foolish man ! that the 
 best thing he could possibly do for his dear child would be 
 to give her the immensest pile of yellow, glistening coin, 
 that had ever been heaped together since the world was 
 made. Thus, he gave all his thoughts and all his time to 
 this one purpose. If ever he happened to gaze for an in- 
 stant at the gold-tinted clouds of sunset, he wished that 
 they were real gold, and that they could be squeezed safely 
 into his strong box. When little Marygold ran to meet him, 
 with a bunch of buttercups and dandelions, he used to say, 
 "Poh, poh, child! If these flowers were as golden as they 
 look, they would be worth the plucking!" 
 
 And yet, in his earlier days, before he was so entirely 
 possessed of this insane desire for riches, King Midas had 
 shown a great taste for flowers.
 
 Reading Scales 85 
 
 No. 6 
 
 In a secluded and mountainous part of Stiria there was 
 in old times a valley of the most surprising and luxuriant 
 fertility. It was surrounded on all sides by steep and rocky 
 mountains, rising into peaks which were always covered 
 with snow, and from which a number of torrents descended 
 in constant cataracts. One of these fell westward over the 
 face of a crag so high that, when the sun had set to every- 
 thing else, and all below was darkness, his beams still shone 
 full upon this waterfall, so that it looked like a shower of 
 gold. It was, therefore, called by the people of the neigh- 
 borhood, the Golden River. It was strange that none of 
 these streams fell into the valley itself. They all descended 
 on the other side of the mountains, and wound away through 
 broad plains and past populous cities. But the clouds were 
 drawn so constantly to the snowy hills, and rested so softly 
 in the circular hollow, that in time of drought and heat, 
 when all the country round was burnt up, there was still 
 rain in the little valley ; and its crops were so heavy and its 
 hay so high, and its apples so red, and its grapes so blue, 
 and its wine so rich, and its honey so sweet, that it was a 
 marvel to every one who beheld it, and was commonly called 
 the Treasure Valley. 
 
 The whole of this little valley belonged to three brothers 
 called Schwartz, Hans and Gluck. Schwartz and Hans, the 
 two elder brothers, were very ugly men, with overhanging 
 eyebrows and small, dull eyes.
 
 86 Scientific Measurement 
 
 No. 7 
 
 Captain John Hull was the mint-master of Massachusetts, 
 and coined all the money that was made there. This was a 
 new line of business, for in the earlier days of the colony 
 the current coinage consisted of gold and silver money of 
 England, Portugal, and Spain. These coins being scarce, 
 the people were often forced to barter their commodities 
 instead of selling them. 
 
 For instance, if a man wanted to buy a coat, he perhaps 
 exchanged a bearskin for it. If he wished for a barrel of 
 molasses, he might purchase it with a pile of pine boards. 
 Musket-bullets were used instead of farthings. The In- 
 dians had a sort of money called wampum, which was made 
 of clamshells, and this strange sort of specie was likewise 
 taken in payment of debts by the English settlers. Bank- 
 bills had never been heard of. There was not money enough 
 of any kind, in many parts of the country, to pay the salaries 
 of the ministers, so that they sometimes had to take quintals 
 of fish, bushels of corn, or cords of wood instead of silver or 
 
 gold. 
 
 As the people grew more numerous and their trade one 
 with another increased, the want of current money was 
 still more sensibly felt. To supply the demand the general 
 court passed a law for establishing a coinage of shillings, 
 sixpences, and threepences. Captain John Hull was ap- 
 pointed to manufacture this money, and was to have about 
 one shilling out of every twenty to pay him for the trouble 
 of making them.
 
 Reading Scales 87 
 
 No. 8 
 
 The years went on, and Ernest ceased to be a boy. He 
 had grown to be a young man now. He attracted little 
 notice from the other inhabitants of the valley; for they 
 saw nothing remarkable in his way of life, save that, when 
 the labor of the day was over, he still loved to go apart and 
 gaze and meditate upon the Great Stone Face. According 
 to their idea of the matter, it was a folly, indeed, but par- 
 donable, inasmuch as Ernest was industrious, kind, and 
 neighborly, and neglected no duty for the sake of indulging 
 this idle habit. They knew not that the Great Stone Face 
 had become a teacher to him, and that the sentiment which 
 was expressed in it would enlarge the young man's heart, 
 and fill it with wider and deeper sympathies than other 
 hearts. They knew not that thence would come a better 
 wisdom than could be learned from books, and a better life 
 than could be molded on the defaced example of other 
 human lives. Neither did Ernest know that the thoughts 
 and affections which came to him so naturally, in the fields 
 and at the fireside, and wherever he communed with him- 
 self, were of a higher tone than those which all men shared 
 with him. 
 
 By this time poor Mr. Gathergold was dead and buried ; 
 and the oddest part of the matter was, that his wealth, 
 which was the body and spirit of his existence, had disap- 
 peared before his death, leaving nothing of him but a living 
 skeleton, covered over with a wrinkled, yellow skin. Since 
 the melting away of his gold, it had been very generally 
 conceded that there was no such striking resemblance, after 
 all, betwixt the ignoble features of the ruined merchant 
 and that majestic face upon the mountainside.
 
 88 Scientific Measurement 
 
 No. 9 
 
 To an American visiting Europe, the long voyage he has 
 to make is an excellent preparative. The temporary 
 absence of worldly scenes and employments produces a 
 state of mind peculiarly fitted to receive new and vivid 
 impressions. The vast space of waters that separates the 
 hemispheres is like a blank page in existence. There is no 
 gradual transition, by which, as in Europe, the features and 
 population of one country blend almost imperceptibly with 
 those of another. From the moment you lose sight of the 
 land you have left, all is vacancy until you step on the 
 opposite shore, and are launched at once into the bustle 
 and novelties of another world. 
 
 In traveling by land there is a continuity of scene and a 
 connected succession of persons and incidents, that carry on 
 the story of life, and lessen the effect of absence and separa- 
 tion. We drag, it is true, "a lengthening chain," at each 
 remove of our pilgrimage ; but the chain is unbroken : we 
 can trace it back link by link ; and we feel that the last still 
 grapples us to home. But a wide sea voyage severs us at 
 once. It makes us conscious of being cast loose from the 
 secure anchorage of settled life, and sent adrift upon a 
 doubtful world. It interposes a gulf, not merely imaginary, 
 but real, between us and our homes — a gulf subject to 
 tempest, and fear, and uncertainty, rendering distance pal- 
 pable, and return precarious. 
 
 The tests and standard scores in this section are reproduced by the 
 courtesy of Dr. Daniel Starch.
 
 Reading Scales 89 
 
 DIRECTIONS FOR ADMINISTERING TESTS 
 
 The series of tests published by Starch are designed to 
 measure (1) comprehension of material read, (2) speed of 
 reading, and (3) correctness of pronunciation. These 
 tests, nine in number, are actually a graded series of 
 passages chosen from various graded readers, each of 
 them bearing a number which indicates the grade from 
 which it was taken and in which it is to be used. For 
 example, No. 1 is to be used in the first grade ; No. 2, in 
 the second grade; and so on. It should be noted that 
 full directions for administering the tests accompany 
 them. 
 
 Explain to the pupils that they are to read silently as 
 rapidly as they can and at the same time to grasp as much 
 as they can, and that they will be asked to write down, 
 not necessarily in the same words, as much as they will 
 remember of what they read. 
 
 They should also be told not to read anything over, 
 but to read on continuously as rapidly as is consistent 
 with grasping what they read. 
 
 Use for a given grade the test blank that bears the 
 same number as that grade. For example, use No. 4 
 with the fourth grade, No. 5 with the fifth grade, etc. 
 On the next day repeat the test in the same manner, but 
 use the blank of the grade next below yours ; that is, in 
 the fourth grade use No. 3, in the fifth grade use No. 4, etc. 
 
 The blanks for the test should be distributed to the 
 pupils with the backs of the blanks up, so that no one will 
 be able to read any of the material until all are ready. 
 Then give the signal "turn" and "start." Allow them 
 to read exactly thirty seconds. Then have the pupils 
 make a mark with pencil after the last word read to indi- 
 cate how far they had read. 
 
 Then have them turn the blanks over immediately and 
 write on the back all that they remember having read.
 
 90 Scientific Measurement 
 
 Allow as much time as they need, but make sure that 
 they do not copy from one another, or turn the blank over 
 to see the text. Finally, have them fill out the spaces at 
 the bottom of the blank. 
 
 Make sure of allowing exactly 30 seconds for the read- 
 ing. See that all pupils start and stop at the same time. 
 
 Since selection No. 1 was taken from a typical first 
 grade reader, selection No. 2 from a typical second grade 
 reader, and so on, it was assumed that the increase in 
 difficulty from one passage to another was fairly uniform. 
 Nevertheless, Starch carefully examined all the data 
 obtained from administering the tests to about 1400 
 pupils. These data indicated that the assumption was 
 correct that the passages increase in difficulty with ap- 
 proximate uniformity from step to step. They also 
 seemed to show that, unless the selections have been read 
 shortly before by the pupils tested, the value of the tests 
 is not affected by the fact that some of the selections are 
 more or less familiar fables or pieces of literature. 
 
 (1) Reading Comprehension Test 
 
 In using the test to measure reading comprehension, 
 the pupil is given a limited time — thirty seconds — to 
 read as much as he can of the selection. He is then re- 
 quired to write out as much as he can of what he has read. 
 The exact amount of understanding shown is determined 
 by counting the number of words written which correctly 
 express the thought of the selection. All words which 
 reproduce the ideas of the test passage incorrectly, all 
 words expressing added ideas or repeated ideas, are 
 crossed out, and the number of remaining words is 
 reckoned as the measure of comprehension. For instance, 
 if a pupil in reproducing test No. 8, a selection of 142 
 words, uses 77 words and 5 are crossed out, his score is 72. 
 
 Starch answers the objection to written reproduction 
 as an index of comprehension, by saying that if it is a
 
 Reading Scales 91 
 
 handicap it is the same for all, since the pupil who is at a 
 distinct disadvantage in writing, as compared with speak- 
 ing, is either very rare or fictitious. Immediate repro- 
 duction was thought best because it does away with the 
 memory factor and imposes uniformity. The immediate 
 memory span of an adult in verbatim reproduction of 
 words in sentences is 25 words, and that of a child of 
 six about 12 words, but in the time allotted for the test, 
 the average eighth grade child will read 120 words and 
 the average first grade pupil will read 45 words. There- 
 fore, the chance of a child's memorizing a great part of 
 the passage is eliminated by the length of the selection. 
 
 Another possible way of testing comprehension is to 
 measure the ability of a child to answer certain questions 
 concerning the test passage. This method was actually 
 tried, but the results from its use were less accurate and 
 more difficult to score than those from the method finally 
 adopted. 
 
 The method of scoring comprehension by counting the 
 number of words written which correctly reproduce the 
 thought of the test passage, was adopted because "it is 
 " simple, rapid and objective." Two other methods, 
 that of assigning percentage marks and that of finding 
 the number of ideas correctly expressed, might have been 
 used ; but the former was disregarded because of its sub- 
 jective character, and the latter, because it involved the 
 difficulty of determining just what an idea is. For in- 
 stance, is " hurried " a separate idea, or should " hurried 
 along " be considered as one? 
 
 (2) Test for Speed of Reading 
 
 The speed of reading is easily measured by determining 
 how much of a given test passage the child is able to 
 read in thirty seconds. By using a blank on which the 
 number of words is indicated line by line to the end of 
 the passage, the total number of words read in a given
 
 92 Scientific Measurement 
 
 time may be seen at a glance. This number, divided by- 
 thirty, is the child's score per second. Thirty seconds 
 was chosen as the time limit, first, because " the neces- 
 sary text for this interval could all be printed on a sheet 
 of paper about the size of an ordinary page in a reader; 
 and second, because a longer interval of time would in- 
 crease very materially the labor of scoring the results." 
 To ascertain whether thirty seconds is a long enough 
 interval to test a pupil's reading capacity, preliminary 
 tests were made which showed that both speed and com- 
 prehension remain nearly constant, irrespective of the 
 length of the passage. 
 
 (3) Test for Correctness of Pronunciation 
 
 Correctness of pronunciation is measured by noting 
 the number of words pronounced incorrectly. The test 
 is administered after the other two tests are completed, 
 and when the pupil has, in consequence, acquired a certain 
 familiarity with the passage. Of course, this test must 
 be given individually and out of the hearing of the other 
 pupils. 
 
 To test the validity of these measurements of reading 
 capacity, a comparison was made in a school of 256 pupils 
 between efficiency in reading as shown by the tests and 
 as indicated by marks in reading assigned by teachers. 
 The relation between the results of the tests and the read- 
 ing as estimated by the teachers was close. 
 
 STANDARD SCORES IN READING 
 
 On the basis of the results from the administration of 
 the tests to over 3500 children in 15 schools in 7 cities 
 and 3 states, the following tentative standard scores of 
 efficiency have been made for each grade.
 
 Reading Scales 
 
 93 
 
 Grades 
 
 Speed 
 
 op Reading (words per second) 
 
 Comprehension (words written) 
 
 1 
 
 
 1.5 
 
 15 
 
 2 
 
 
 1.8 
 
 20 
 
 3 
 
 
 2.1 
 
 24 
 
 4 
 
 
 2.4 
 
 28 
 
 5 
 
 
 2.8 
 
 33 
 
 6 
 
 
 3.2 
 
 38 
 
 7 
 
 
 3.6 
 
 45 
 
 8 
 
 
 4.0 
 
 50 
 
 These tests show that great individual differences exist 
 among pupils in the same grade. For example, in one 
 of the fourth grades tested, one pupil showed a speed of 
 reading of .8 words per second and another, of 4.7. Since 
 the standard for speed in the first grade is 1.5 words per 
 second and in the eighth grade, 4.0 words per second, the 
 former pupil falls considerably below the standard of the 
 first grade and the latter rises above the standard of the 
 eighth grade. The same holds true of comprehension. 
 In combined speed and comprehension the best pupil in 
 one fourth grade made a score four and one half times 
 as high as the poorest. 
 
 This wide difference in ability in a single grade means 
 that a large amount of overlapping exists between dif- 
 ferent grades. In fact, on the basis of the studies made 
 so far with these tests, it may be said that " one third of 
 the pupils of any given grade could do the reading work 
 of the next grade above as well as the average of that 
 grade, one fifth could do the work of the second grade 
 above as well as the average of that grade, and one 
 eighth could do the work of the third grade above as well 
 as the average of that grade."
 
 94 Scientific Measurement 
 
 HI. COURTIS READING TESTS 
 
 Courtis has constructed two different reading tests, 
 one to measure rate and retention in normal reading 
 (Test No. 4, Normal Reading, Series C) and the other to 
 measure rate and retention in careful reading (Test No. 5, 
 Careful Reading, Series C). 
 
 Rate of normal reading is determined by telling the child 
 to read a selected passage for one minute at his natural 
 rate of reading. (This test will be found reproduced on 
 pages 96 and 97.) 
 
 At the end of this time the pupil is to draw a circle 
 around the last word he has read. Since the words are 
 numbered to the end of the passage his rate of reading 
 may be quickly determined. 
 
 Retention in normal reading is measured by giving the 
 pupil a sheet of paper on which is the story that he has 
 just read, but having in it here and there groups of three 
 words (in parentheses), two of which words were not 
 used in the original story. (See pages 98 and 99.) 
 
 He is to cross out the words which he does not re- 
 member seeing before, and, if he is unable to recall whether 
 he has seen them or not, he is to cross them all out. He 
 is to continue this until he comes to the word at which 
 he stopped in the original story. For remembering to 
 stop at this place he is given credit for one point. 
 
 Scoring is done by means of an Answer Card (see 
 page 104) which gives the correct words used in the orig- 
 inal story. This is placed beside the pupil's paper and 
 every word which has been correctly crossed off is counted 
 as a point. By adding these points the pupil's score in 
 retention is obtained. 
 
 The rate of careful reading is determined in a similar 
 manner. Retention, however, is measured by the amount 
 of the selected passage that the child is able to reproduce. 
 (See tests on pages 100 and 101.)
 
 Reading Scales 95 
 
 Scoring is done by means of an Answer Card (see 
 pages 102 and 103) which contains a list of the points or 
 main ideas in the passage. For the reproduction of each 
 of these ideas the pupil is given a credit of one. His 
 final score is the sum of these credits. 
 
 The following tests (pages 96 to 104) are reproduced by the cour- 
 tesy of Mr. S. A. Courtis.
 
 96 
 
 Scientific Measurement 
 
 £ 
 
 
 
 
 O 
 
 o 
 
 
 
 U 
 
 
 
 
 co 
 
 
 CO 
 
 
 3 
 
 O 
 
 
 z 
 
 e* 
 
 o -o 
 
 . >- 
 
 o o 
 Z^ 
 
 a 
 •a 
 a> 
 
 a 
 
 o 
 
 o 
 
 CO 
 
 H 
 
 .d 
 w 
 
 *bb 
 
 w 
 
 13 
 0< 
 
 2> 
 
 o 
 
 CO 
 
 o 
 C 
 
 00 
 
 00 cm t>. © co >© 
 
 1-4 Co * ie i« io 
 
 » CO oo <n \o 
 a ih n ^ in 
 
 CO oo CM 
 oo o* <H 
 
 3 - cj ,d CO 
 
 "T o3 M T3 
 
 +2 £ - ® 
 
 2P - 73 bfl bfi 
 
 _2 2 M o 
 
 O CO . 
 
 bc9 H 
 
 .flS 
 
 CO 
 
 ,d 
 d 
 
 ^3 
 
 d 
 
 o 
 
 b€ 
 O 
 
 in 
 
 d 
 
 CO 
 
 > 
 co 
 
 o3 
 
 Ph 5- d O _ 
 
 .d +J 
 J-c 
 
 <1> 
 
 d 
 
 T3 
 d 
 03 
 
 d 
 
 d 
 
 8 * 
 
 co 
 
 si 
 
 o3 
 
 ^s a 
 
 in 
 
 CO 
 
 > «j co £ -^ 
 
 o « ^ b ° a 
 
 — « co ^ =: J5 
 
 S-» CO ^ -"(J ■— ' 
 
 • »— « • r/} ** <»r-t k. 
 
 W) bJO - fl >? 
 
 co -p S 1^ 
 
 +3 o d ^ „, bo 
 
 -m R o3 co 03 
 
 ^ -^ n^ co rd co 
 
 s 
 
 T5 § ^ 
 
 o3 
 
 §* 
 
 co bo .. wm 
 
 £-° 03^3 
 ^ *3 ^ n 
 
 *•"' 73 X! PI 
 
 en be .. a 
 T3 T3 »" eS 
 
 b€ C 
 
 55 .-a 
 
 5-i .i— i 
 
 .9 X 
 
 73 
 
 73 ^ ^ 
 
 s 5 ° 
 
 ^ 2 S e2 bb £? 
 PQ 2 
 
 +5 s- o 
 d f3 cu m 
 
 o ^ -d ^S 
 
 73 
 O 
 
 d 
 
 .2 ° 73" 
 
 -M ^^ 'O 
 
 .a bo d 
 -. g « 
 
 -d C^ 
 
 +* ** «M 
 
 T3T! ° 
 CO c3 
 
 ^ S^ 
 P3 ,__ CO 
 
 C 73 en 
 
 •S £ C M ^3 
 
 73 TO TO CO 
 
 «i-H fS- 
 d g ^ r-. 
 
 * +» ^ •§, 
 ■ +J T3 T3 -M 
 
 % *£ §3 * I 
 
 4J t> ^ O i> 
 
 CO 
 
 o 1 
 
 CO ^ 
 
 CO 
 
 73 
 $-1 
 
 o 
 o 
 
 CO 
 
 -3 o 
 
 bO 
 
 03 
 
 So & 
 
 bo 
 
 o 
 
 >> CO 
 
 d d 
 
 o CO CO 
 
 d ^-' 
 
 CO 
 
 co d 
 
 CO 03 ^ 
 
 "S 2 *h 
 
 o o 
 
 S M ^S 
 
 o 3 g co a 5 _. 
 
 4J «■, d ^5 d ™ 
 
 «„ d 'rt -P g ^ 
 
 ^ +» "8 „, .5 co 
 
 S d I c3 s 1 
 
 8 co^ a 
 
 1 a 88 « 
 
 be rn 03 d 
 
 2 a> -S -P 
 
 d ^ -C bo "^ hn 
 
 d -> *« "So ^ o 
 
 cy bo O $ £ *=! 
 
 T3 d xn A r „ c3 
 m 'd a> o ^ 
 
 be pj t» 73 co O 
 
 fl ^ co ^3 03 x «w 
 9 S-i J2 S -M 
 
 O o 
 
 ■a 'S ^ -s « 
 
 o o3 bo F c3 • 
 
 j2 d ^ SB 
 
 03 bp o bO d $
 
 Reading Scales 
 
 97 
 
 en 
 
 0\ — 
 
 CO 
 
 o * 
 
 o *4 
 ro to 
 
 
 m 
 
 to 
 
 CO 
 
 oo 
 <o 
 
 03 
 
 l-c 
 
 o 
 
 4} 
 
 a 
 
 at 
 
 £ o &pq p,
 
 98 
 
 Scientific Measurement 
 
 w 
 
 
 
 
 rt 
 
 
 
 
 o 
 
 
 
 
 o 
 
 
 13 
 
 
 CO 
 
 
 (0 
 
 00 
 
 
 V 
 
 t- 
 
 C 
 
 
 E 
 
 d 
 
 d 
 
 
 E- 
 
 Z 
 
 «i 
 
 •h m « m io 
 
 "8 
 
 •5 
 
 c 
 
 a 
 
 (30 
 
 a 
 
 a> 
 
 »-^ 
 
 as 
 
 a 
 
 i-i 
 
 o 
 
 I 
 
 ■<* 
 6 
 
 <u 
 H 
 
 -4-> 
 
 B 1 
 
 © 
 
 CO 
 
 CO* 
 
 c 
 <J> 
 => 
 ■a 
 
 co* 
 
 o 
 C 
 
 ■ — 
 
 CO 
 
 a> 
 
 • •— 
 co 
 
 CO 
 
 oi © >-< e* w> m <o 
 
 co en o 
 
 8 
 •<s> 
 
 "<3 
 
 •<s> 
 
 «J 
 
 ■to 
 
 ""3 
 
 
 o 
 © 
 
 13 
 S 
 
 CO 
 
 s 
 
 •AS 
 
 13 
 
 s 
 
 c 
 
 03 
 
 bO 
 
 cu 
 
 Xi 
 
 -d 
 
 i 
 
 03 
 
 «* , 
 
 be: 
 
 03 . 
 
 .Si 
 m . 
 
 m '■ 
 O 
 
 a 
 +j 
 
 a3 
 
 W 
 
 U) 
 
 O 
 
 X5 
 
 <D 
 
 2 
 
 -4-3 
 
 -t-J 
 
 ©• 
 
 "J 
 
 cu 
 '3 & 
 
 m 
 
 03 
 O CU 
 
 43 *-. 
 
 C CD 
 
 
 cu 
 
 .x5 : 
 5 
 
 03 
 
 *< 03 
 
 00 cu 
 
 e ft 
 
 J ° 
 
 . o 
 
 4— • 
 
 a 
 .Si 
 
 t3 
 
 o3 
 
 s 
 
 bo 
 
 +j 
 
 
 
 CU 
 
 
 *-*s 
 
 -O 
 
 £ 
 
 O 
 
 <v-t 
 
 •1— t 
 
 t/j 
 
 CO 
 
 S-4 
 
 cu 
 
 Fi 
 
 
 -C 
 
 a) 
 
 to 
 
 4-> 
 
 
 
 0) 
 
 "■3 
 
 q 
 
 
 
 03 
 
 +J 
 
 >~. 
 
 0) 
 
 cu 
 J-l 
 o 
 
 cu 
 
 PQ 
 
 4J 
 
 03 
 © 
 
 CU 
 
 > 
 
 03 
 cu 
 
 C35 
 
 © 
 i- 
 
 X3 „- 
 
 o3 
 X2 
 
 09 
 
 CU
 
 Reading Scales 
 
 99
 
 100 
 
 Scientific Measurement 
 
 § 
 
 
 
 
 o 
 
 o 
 
 
 
 O 
 
 
 
 
 CO 
 
 
 CO 
 
 1m 
 
 
 3 
 
 O 
 
 
 z 
 
 £ 
 
 be 
 d 
 
 cS 
 
 o 
 
 10 
 
 6 
 
 CO 
 
 CD 
 
 H 
 
 J3 
 
 CO 
 
 • «-« 
 
 "So 
 
 c 
 W
 
 Reading Scales 
 
 101 
 
 Ol r* m t» r-t 
 
 t* o> o <-< o 
 
 •H r4 C4 O N 
 
 
 h .-B x! -c
 
 102 
 
 Scientific Measurement 
 
 r* N CO 
 
 i/) <o 
 
 c£ 
 
 bO 
 
 a 
 
 Pi 
 
 crj 
 
 3 
 
 co 
 
 u 
 
 a 
 o 
 
 
 > 
 
 
 "O 
 
 in 
 
 <! 
 
 
 CO 
 
 
 
 o 
 
 a) 
 
 fc 
 
 co 
 
 CO 
 
 CO 
 4) 
 
 H 
 
 0)
 
 Reading Scales 
 
 103 
 
 s S §>-§ 
 
 
 O 
 Ph 
 
 0) 
 
 <-« CO 
 
 S.2 
 t-, 
 
 +j en 
 
 c ra e3 
 
 2 "S 
 
 S -Q « 
 
 £ 
 
 T3 
 CD 
 
 S 
 
 "C 
 
 OJ 
 
 +5 
 
 0) 
 
 
 c( S g 
 
 
 pq £ 
 
 a> 
 
 o 
 
 S 
 
 > 
 
 o *- 
 o 
 
 Ih 
 
 «-. 
 CO (1) 
 
 *© <c> t^ 00 as 
 
 2 ^ 
 
 £ 
 
 -fi o 
 
 o ° 
 
 fcJO W 
 
 CD O 
 
 '5 £3 
 
 £ s 
 
 T3 
 
 >, >> cd 
 
 c », S 
 
 CD O 
 
 o 
 
 S 
 
 S3 
 
 o 
 CD 
 -Q 
 
 3 -fi JB 
 < O M 
 
 a> 
 ci 
 
 -M .5 
 
 3 S3 
 
 F-h CD 
 
 o 2 
 
 60 
 
 S3 
 
 o 
 
 CD 
 S3 
 
 o 
 u 
 
 CD 
 
 T3 O 
 
 cd a> 
 
 S3 £ 
 
 EH 
 
 51 O HN W 1» 
 
 a) 
 
 S3 
 >-. 
 
 .2 
 '55 
 
 CD 
 
 PQ 
 
 CD 
 S3 
 I- 
 O 
 u 
 
 o 
 t-l 
 
 J3 
 
 T3 
 CD T3 
 
 2 * 
 
 60 
 P S3 
 
 J 03 tf 
 
 CD 
 
 C CD 
 
 J3 *■" 
 60 _g 
 
 <4-H 'P 
 
 ° s 
 
 O CD 
 
 a « 
 
 60 
 O 
 
 CD 
 
 u 
 
 CO CD 
 X S3 
 
 o o 
 
 S £ 
 2 ^ 
 
 £.2 
 
 to S 
 
 Q PQ 
 
 HNW^IOIOC-OO
 
 104 
 
 Scientific Measurement 
 
 9 
 
 a 
 <o 
 
 a 
 
 O 
 
 -J rid 
 
 O w 
 
 t> rQ 
 
 rG G . ^ 
 
 "« » a "- 
 
 O O — i »-i 
 
 cd 
 
 • 
 
 G co 
 O bO 
 
 *S .3 
 
 03 i t3 
 - G ^ 
 
 * a - 
 
 SC CD O Cfl 
 
 1 6b ll 
 
 -G tT O 
 
 °i & 
 
 S* TJ _jh «+-i G> 
 
 a .g •§ (£ & j* 
 
 PS to Q H A ft 
 
 G 
 
 >» G £ 
 
 S a s 
 
 co G ^ 
 
 _ "2 •» 
 
 « Oi (y 
 TTJ Pi "H 
 
 CD 
 
 0) 
 
 rG 
 
 rG 
 >> 
 
 03 
 <x> 
 
 -e 
 
 • 9? 
 -b ^ 
 
 S. ™ 
 
 53 o 
 
 O «4-t 
 
 CU CU 
 
 to pq 
 
 G -G 
 
 £ rG 
 
 ft 
 
 G 
 O 
 
 co •> 
 bfl G 
 
 .as 
 
 G cu 
 
 co +-> 
 G - 
 
 G 
 G 
 ft 
 
 r* 
 
 o 
 
 • r-t 
 
 G 
 M 03 
 
 to 
 
 .a 
 
 G 
 G 
 o 
 
 • G 
 
 a> .G 
 
 S-i >— i 
 
 03 
 
 G 
 
 a 1 
 
 co 
 
 
 G 
 G 
 5 
 
 co 
 
 CO 
 
 cu 
 
 w w ft ib fe w Q 
 
 cu o 
 
 
 W 
 +■» 
 S 
 Pi 
 
 r»> 
 
 o 
 
 a> 
 
 • r- 1 
 
 45 
 
 O 
 
 H 
 
 rH 
 
 «4-H 
 
 
 
 
 s 
 
 
 c3 
 
 
 o 
 
 2 £ G 
 
 G o3 o3 
 d co p» 
 bO * 
 
 C3 +j 
 Qd O 
 
 •■o 
 
 ^> q^ 
 
 G 
 
 Ej • !— | 
 
 0) 
 
 ri-H 
 
 g So 
 
 fc 
 
 03 CD 
 
 
 
 I 
 
 g 
 
 ■a 
 
 CO 
 
 bO bO 
 G G 
 
 • r-t »r— I 
 
 -^ ^G 
 
 co 
 
 o 
 
 03 
 
 tf V 
 tf 
 
 CO 
 
 03 
 
 a 
 
 03 
 O 
 
 cu 
 
 m 
 
 T3 
 O 
 
 (MW^WCDt>00C5OH(NC0rt<incfit>000>OH(M05 
 NWWNOJNi^(>qcOWMMcOCOCOCOCOCO'<*rt*'^Tj* 
 
 cu 
 O 
 
 CQ 
 
 <u 
 
 S3 
 O 
 
 CD 
 
 C 
 
 G lL — 
 p -G w c« 
 
 I 
 
 O 
 
 bO 
 
 S=S 
 •i-h 
 
 42 
 
 CD 
 
 i 
 
 o 
 
 o 
 o 
 
 O 
 
 CO 
 
 cu 
 H 
 
 O 
 co 
 
 co 
 
 CD 
 ti 
 
 G 
 +j 
 
 G 
 o 
 > 
 
 •G 
 
 o 
 
 iS 
 
 o ^ 
 
 ft ^ 
 
 • G 
 
 __- • r— I 
 
 O 5 
 
 G . 
 o oT 
 
 -G :2 
 
 bO o 
 *C « 
 
 fa Q 
 
 0) 
 
 G 
 
 .G 
 
 rM 
 
 rH 
 
 03 
 
 rQ 
 
 CO 
 
 rH 
 
 co 
 
 rG 
 
 co 
 
 rl 
 
 c3 
 
 rG 
 
 rb 
 
 ^3 
 
 O 
 rQ 
 
 
 r. d. CO 
 
 rn 
 O 
 
 b 
 
 -is 
 
 >> c3 
 
 3 ^ 
 
 a 
 
 G 
 
 cr 
 
 03 
 
 CD •» • r. 
 
 jj 0) G .G a> S 
 
 • • 
 
 G ° 
 
 3 G 
 
 bOr^T 
 
 ex <^ 
 
 fl O « 3 . 
 
 o o a> 5* y 
 
 co 
 CD 
 
 03 
 
 r* 
 
 ^ bO 
 
 bO G 
 
 .a -a 
 
 ^ ft 
 
 c3 
 'ft 
 
 bO 
 
 G 
 
 M " G G 
 
 d> g 0J 
 
 • r4 n« »H 
 
 fa O CO 
 
 csj 
 
 8 G 
 
 rG .a 
 
 X3 w 3 +3 
 
 HNM^WCOt-OOOJOH NcO^lOOt^OOOJ 
 
 T3 
 CD 
 
 CO 
 
 CO 
 
 G .T3 
 
 CD r, 
 
 r" *-" 
 
 rH CD 
 
 I 
 
 CU 
 
 ol 
 PQ 
 
 G 
 
 c3 cd 
 
 ^ I 
 
 G co 
 +j 03 
 
 -73 CD O 
 
 CD 
 
 o 
 
 13 
 
 O co co rj 
 
 ri2 -2 rg I 
 
 °S ^ c3 c3 
 
 •> ^ rH ■— ' 
 
 h E* 8 «r 
 
 L. 2 H IB 
 
 55 ra o 'S 
 
 > ^ O G 
 
 G 
 
 I 
 
 G _o 
 CD '" 
 
 rQ - 
 
 r. M 
 
 co G 
 
 * "Si 
 
 •^ G 
 co G 
 
 G G 
 
 PQ »-» 
 
 o 
 a 
 
 o 
 
 r4 
 
 a> 
 
 rO 
 
 3 
 *c3 
 
 * w - ft 
 
 co bO 
 
 CD 
 
 CD 
 
 G 
 
 i-H 
 
 H 
 
 03 
 
 CD 
 
 T3 
 
 rf 
 
 
 G 
 
 O 
 
 
 a 1 
 
 +J 
 
 
 CO 
 
 CU 
 OQ 
 
 PQ
 
 Reading Scales 105 
 
 To summarize then, a teacher of the fourth, fifth, sixth, 
 seventh or eighth grade may test the ability of pupils, 
 (1) in the understanding of single words by using Thorn- 
 dike's Scale A, (2) in the comprehension of material read, 
 by using Thorndike's Scale Alpha, the Starch series of 
 tests, or the Courtis tests, and (3) in the rate of reading, 
 by using either the Starch or the Courtis tests, preferably 
 the Starch. 
 
 Thorndike's Scales may be obtained by sending to 
 Teachers College, Columbia University, New York, and 
 the Starch Reading Tests, by sending to the author at 
 the University of Wisconsin. The sheets on which the 
 scales and tests appear contain full directions for their use. 
 
 In using Scale A the teacher should allow thirty minutes 
 for the test in the fourth grade, twenty-five minutes in 
 the fifth and sixth grades, and twenty minutes in the 
 seventh and eighth grades. In administering Scale Alpha 
 the teacher should allow from twenty to thirty minutes. 
 In Scale A the pupil's score is the highest numbered line 
 that he marks correctly without more than a single error. 
 Scale Alpha is scored in a similar manner ; that is, the 
 pupil's score is the highest numbered step or set in which 
 he has answered at least three of the four questions 
 correctly. 
 
 In using the Starch Tests the teacher should send for 
 the test blank that bears the same number as her grade ; 
 for example, No. 4 for the fourth grade, No. 5 for the 
 fifth grade, etc. The speed of reading is obtained by 
 determining the number of words read in thirty seconds. 
 The pupil's score is determined by counting the number 
 of words in his written reproduction which correctly 
 express the thought of the selection read. Added and 
 repeated words, as well as those which represent the ideas 
 of the selection incorrectly, are not counted. 
 
 Folders or manuals, covering every phase of the test- 
 ing, together with answer cards, must be procured with
 
 106 Scientific Measurement 
 
 the test sheets if the Courtis Tests are to be used. These 
 may be obtained by sending to the Department of Co- 
 operative Research, 82 Eliot Street, Detroit, Michigan. 
 To measure oral reading Gray's Scale may be used. 
 In reading the paragraphs in the scale, which gradually 
 increase in difficulty, the gross errors, minor errors, omis- 
 sions, substitutions, and insertions made in each para- 
 graph are recorded. If a child makes 4 or more errors 
 in a paragraph and takes 30 seconds or more to read it, 
 or if he makes 5 or more errors, however quickly he reads, 
 he may be considered to have failed in that paragraph. 
 This scale may be obtained by sending to Teachers Col- 
 lege, Columbia University, New York City. 
 
 EXERCISES 
 
 1. Describe in detail the methods you would employ for measur- 
 ing the reading ability, oral and silent, of thirty children of Grade V, 
 using (a) the Thorndike and Gray Scale and (6) the Starch Scale. 
 
 2. How would you compare your class with one of the same grade 
 in another school, using the Starch Scale? What conditions would 
 you have to meet to make the comparison of the results valid? 
 
 3. How do the results obtained from the Thorndike Scale compare 
 with those which the Starch Scale give? 
 
 4. Does there seem to be any relation between speed of reading 
 and comprehension of material read? 
 
 5. What distinctions between oral and silent reading have the 
 tests revealed? 
 
 6. Have the tests revealed any marked difference in the reading 
 ability of boys and girls? Of children of different nationalities? Of 
 children who have used different reading textbooks? 
 
 7. In what way may a teacher modify Scale A so as to use it to 
 test knowledge in various subjects in the curriculum from the ele- 
 mentary grades through college? 
 
 8. When should a teacher stop drill in oral reading and devote all 
 the time to drill in comprehension? 
 
 9. Have the tests revealed wide variations in the reading ability 
 of the pupils in your class or a condition of more or less uniformity? 
 
 10. What are the shortcomings of the scales described in this 
 chapter? How could these be remedied?
 
 CHAPTER V 
 SPELLING SCALES 
 
 I. BUCKINGHAM SCALE 
 II. STARCH SCALE 
 HI. AYRES SCALE 
 
 I. BUCKINGHAM SPELLING SCALE 
 
 This investigation, following the lead taken by the ex- 
 perimental investigation of the quality of handwriting and 
 of composition, had as its object the development of a scale 
 for the measurement of spelling ability; a scale which 
 would no longer depend upon chance selection of words 
 and upon subjective judgments of teachers, but which 
 would be of general application and purely objective. The 
 results were first published in 1913. 
 
 It is an obvious fact that there is a great difference in 
 words as regards ease of spelling. Thus, we can select 
 words of the very simplest, such as the, as, when, up to 
 words of extreme difficulty which can only be spelled after 
 long acquaintance. Theoretically, therefore, it is possible 
 to arrange a series of words along a scale in such a way that 
 they become more and more difficult. Furthermore, it 
 might be possible to arrange these words at equal inter- 
 vals along the scale, these intervals being determined by 
 the difficulty of each word. If in addition to this we fix 
 a zero point (by taking the simplest words and agreeing 
 that failure to spell these words indicates absence of spell- 
 ing ability), a scale may be constructed which will meas- 
 ure the spelling ability of any individual, and will measure 
 
 107
 
 108 Scientific Measurement 
 
 the difficulty of any word which has to be spelled. Not 
 only can we measure the spelling ability of individuals in 
 this way, but also of classes, schools, and school systems. 
 
 Such measurements will be independent of individual 
 opinion. Spelling ability will be determined, not by an 
 arbitrary list of words, picked at random by individuals 
 who have no knowledge of their relative degrees of diffi- 
 culty, but by means of words on the scale, which have 
 been standardized as regards their difficulty, by the simple 
 device of finding out what percentage of eighth grade 
 children spelled them correctly. The school has always 
 attached great importance to spelling ability; whether 
 or not this ability is overestimated, does not need discus- 
 sion here. Suffice it to say, that if the school takes as its 
 aim the teaching of spelling, it is essential that some 
 method be devised to measure the extent to which the 
 aim is accomplished. 
 
 Dr. Rice, as early as 1897, tested the pupils in all grades 
 from the fourth to the eighth inclusive in twenty-one 
 school systems, using a list of words, which has since 
 become known as the Rice Sentence Test. This list is 
 given on the following page.
 
 Spelling Scales 
 
 109 
 
 
 
 RICE 
 
 SENTENCE 
 
 LIST 
 
 
 1. 
 
 running 
 
 30. 
 
 writing 
 
 59. 
 
 sensible 
 
 2. 
 
 slipped 
 
 31. 
 
 language 
 
 60. 
 
 business 
 
 3. 
 
 listened 
 
 32. 
 
 careful 
 
 61. 
 
 answer 
 
 4. 
 
 queer 
 
 33. 
 
 enough 
 
 62. 
 
 sweeping 
 
 5. 
 
 speech 
 
 34. 
 
 necessary 
 
 63. 
 
 properly 
 
 6. 
 
 believe 
 
 35. 
 
 waiting 
 
 64. 
 
 improvement 
 
 7. 
 
 weather 
 
 36. 
 
 disappoint 
 
 65. 
 
 fatiguing 
 
 8. 
 
 changeable 
 
 37. 
 
 often 
 
 66. 
 
 anxious 
 
 9. 
 
 whistling 
 
 38. 
 
 covered 
 
 67. 
 
 appreciate 
 
 10. 
 
 frightened 
 
 39. 
 
 mixture 
 
 68. 
 
 assure 
 
 11. 
 
 always 
 
 40. 
 
 getting 
 
 69. 
 
 imagine 
 
 12. 
 
 changing 
 
 41. 
 
 better 
 
 70. 
 
 peculiar 
 
 13. 
 
 chain 
 
 42. 
 
 feather 
 
 71. 
 
 character 
 
 14. 
 
 loose 
 
 43. 
 
 light 
 
 72. 
 
 guarantee 
 
 15. 
 
 baking 
 
 44. 
 
 deceive 
 
 73. 
 
 approval 
 
 16. 
 
 piece 
 
 45. 
 
 driving 
 
 74. 
 
 intelligent 
 
 17. 
 
 receive 
 
 46. 
 
 surface 
 
 75. 
 
 experience 
 
 18. 
 
 laughter 
 
 47. 
 
 rough 
 
 76. 
 
 delicious 
 
 19. 
 
 distance 
 
 48. 
 
 smooth 
 
 77. 
 
 realize 
 
 20. 
 
 choose 
 
 49. 
 
 hopping 
 
 78. 
 
 importance 
 
 21. 
 
 strange 
 
 50. 
 
 certainly 
 
 79. 
 
 occasion 
 
 22. 
 
 picture 
 
 51. 
 
 grateful 
 
 80. 
 
 exceptions 
 
 23. 
 
 because 
 
 52. 
 
 elegant 
 
 81. 
 
 thoroughly 
 
 24. 
 
 thought 
 
 53. 
 
 present 
 
 82. 
 
 conscientious 
 
 25. 
 
 purpose 
 
 54. 
 
 patience 
 
 83. 
 
 therefore 
 
 26. 
 
 learn 
 
 55. 
 
 succeed 
 
 84. 
 
 ascending 
 
 27. 
 
 lose 
 
 56. 
 
 severe 
 
 85. 
 
 praise 
 
 28. 
 
 almanac 
 
 57. 
 
 accident 
 
 86. 
 
 wholesome 
 
 29. 
 
 neighbor 
 
 58. 
 
 sometimes 
 

 
 110 Scientific Measurement 
 
 The method of scoring was of the simple type which is 
 usually found in schools, i.e. a mark was given for each 
 word correctly spelled, or a unit subtracted for each word 
 misspelled. That is, all words were taken as equal 
 measures of spelling ability. It should be noted that 
 the foregoing list contains among other words, disappoint, 
 necessary, changeable, better, because, picture. An examina- 
 tion of these six words shows at once that they are by 
 no means of equal difficulty. This was conclusively 
 proved by Thorndike, who made an actual test of these 
 words on a group of fifth grade children. Thus, in the 
 group that he measured, while 37% failed to spell neces- 
 sary, the failures to spell better, because, and picture, were 
 3%, 1%, 0%, respectively. This clearly shows that it is 
 erroneous to measure the score of the individual by giving 
 equal value to each of these words. The pupil who scores, 
 let us say, 95%, has spelled not only all the easy words in 
 the list, but also a considerable number of the hard ones, 
 whereas the pupil who gets 50% has failed in the hard 
 words, and has obtained his mark merely by spelling the 
 easy words. That is, as the score increases, the units 
 really get greater and greater, for to spell the five hardest 
 words represents a very different task from spelling the 
 five easiest words, and yet both have the same effect on 
 the score. In other words, studies of this type must 
 always lack precision because of the inequality of the units 
 which are employed. They are useful for giving a rough 
 estimate of the abilities of various groups, but when it 
 comes to asking questions, such as : How does the spell- 
 ing ability of one class differ from another ? — the figures 
 which represent the results give no quantitative informa- 
 tion, and are actually misleading. As the science of 
 school measurement advances, such a state of affairs can 
 hardly be tolerated. Exact quantitative measurements 
 of spelling ability are required ; such quantitative results 
 can never be obtained so long as the fundamental error
 
 Spelling Scales 111 
 
 is made, that one word is equal to another word in 
 difficulty, unless this is proved to be the case by actual 
 measurements of large groups. To correct this error 
 was the purpose of Buckingham's study of spelling 
 ability. 
 
 The study was confined to grades from the third to the 
 eighth, inclusive, of elementary schools located in or near 
 the city of New York. The schools drew such different 
 classes of children that any conclusions derived as a result 
 of the study can be taken as representative. In all, about 
 9000 pupils were tested, a number from which general 
 results might be expected; a greater number of pupils 
 would not have increased the accuracy of the results 
 sufficiently to compensate for the additional labor. 
 
 In the first test a list of 270 words was used. This will 
 be called the "original list." This list was selected from 
 a larger list of 5000 words taken from two or more of 
 five special books used by the author in his own school. 
 These 270 words had to satisfy two requirements : (1) All 
 of them had to be words in the speaking vocabulary of a 
 third grade child, and (2) a considerable portion of the 
 words had to be of sufficient difficulty to test the spelling 
 ability of an eighth grade child. These words were then 
 placed in a continuous passage, and the whole dictated 
 to Grades III to VIII in one school and to Grades IV to 
 VII in another school. The dictation was very slow, so 
 that the time factor did not enter. In marking the papers 
 only the 270 words were regarded, those that served to 
 link the whole into a continuous passage being neglected. 
 All the papers were marked by the same person and two 
 measurements were recorded : (1) the number of times 
 each word was correctly spelled in each grade, and (2) the 
 percentage of the entire number of words each pupil 
 spelled correctly in each grade. We shall confine our- 
 selves to the first consideration, i.e. to the number of times 
 each word was correctly spelled.
 
 112 
 
 Scientific Measurement 
 
 TABLE I 
 Figures Indicate Per Cent Correct 
 Table reads: across was spelled correctly in the third grade of 
 School II by 17% of the pupils; in the fourth grade of School I by 
 60% of the pupils, and of School II by 40% of the pupils, etc. 
 
 Grade .... 
 
 3d 
 
 II 
 
 4th 
 
 5th 
 
 6th 
 
 7th 
 
 8th 
 
 School . . . 
 
 I 
 
 II 
 
 I 
 76 
 
 II 
 58 
 
 I 
 
 II 
 
 I 
 
 II 
 
 II 
 
 across .... 
 
 17 
 
 60 
 
 40i 
 
 90 
 
 79 
 
 98 
 
 87 
 
 93 
 
 addition . . . 
 
 2 
 
 38 
 
 26 
 
 60 
 
 28 
 
 76 
 
 45 
 
 94 
 
 76 
 
 83 
 
 almost .... 
 
 16 
 
 62 
 
 41 
 
 73 
 
 65 
 
 88 
 
 75 
 
 80 
 
 81 
 
 87 
 
 alphabet . . . 
 
 25 
 
 13 
 
 1 
 
 63 
 
 12 
 
 40 
 
 46 
 
 82 
 
 43 
 
 68 
 
 arithmetic . . 
 
 27 
 
 89 
 
 53 
 
 100 
 
 72 
 
 96 
 
 92 
 
 100 
 
 97 
 
 98 
 
 bridge .... 
 
 29 
 
 59 
 
 42 
 
 87 
 
 52 
 
 98 
 
 85 
 
 100 
 
 94 
 
 97 
 
 button .... 
 
 14 
 
 50 
 
 35 
 
 70 
 
 49 
 
 77 
 
 63 
 
 84 
 
 62 
 
 83 
 
 choose .... 
 
 6 
 
 25 
 
 10 
 
 37 
 
 31 
 
 62 
 
 37 
 
 67 
 
 55 
 
 65 
 
 day 
 
 97 
 
 100 
 
 98 
 
 96 
 
 100 
 
 100 
 
 99 
 
 100 
 
 100 
 
 100 
 
 guess .... 
 
 6 
 
 29 
 
 17 
 
 67 
 
 30 
 
 77 
 
 50 
 
 82 
 
 66 
 
 85 
 
 handful .... 
 
 36 
 
 47 
 
 33 
 
 46 
 
 19 
 
 76 
 
 33 
 
 75 
 
 63 
 
 57 
 
 pshaw .... 
 
 1 
 
 4 
 
 6 
 
 29 
 
 6 
 
 46 
 
 5 
 
 31 
 
 31 
 
 18 
 
 tomato . . . 
 
 34 
 
 83 
 
 49 
 
 67 
 
 43 
 
 74 
 
 48 
 
 79 
 
 32 
 
 38 
 
 too 
 
 
 
 10 
 
 3 
 
 17 
 
 4 
 
 26 
 
 7 
 
 63 
 
 22 
 
 27 
 
 whose .... 
 
 17 
 
 49 
 
 15 
 
 40 
 
 29 
 
 47 
 
 10 
 
 57 
 
 59 
 
 66 
 
 Table I represents the typical results obtained from 
 the various grades in the particular schools. Thus for 
 example, across was spelled correctly in the third grade of 
 school II by 17% of the pupils, and in the seventh grade 
 of school I by 98%. On the basis of these scores a group 
 of 100 words, here called the " selected list," was chosen 
 from the original list of 270 words. 
 
 The basis upon which the " selected list " was chosen is 
 as follows : Referring to Table I, it will be seen that the 
 word across was spelled by 17% of the third grade children, 
 which means that it was not too hard to serve as a test of 
 their ability. By the time the seventh and eighth grades
 
 Spelling Scales 113 
 
 were reached, it still served as a test of ability, for it failed to 
 be spelled in the seventh and eighth grades by 13% and 7%, 
 respectively. For this reason the word across was selected. 
 Almost and button were chosen for the same reason. On the 
 other hand, addition, which was spelled by only 2% of the 
 third grade children, was discarded as too difficult, for 2% 
 could spell it rightly by mere chance, which means that 
 the word really serves as no test for the particular grade. 
 
 Continuous Passage — ioo Selected Words 
 
 Whose answer is ninety? If the janitor sweeps, he 
 will raise a dust. You ought not to steal even a penny. 
 Wait until the hour for recess to touch the button. 
 Smoke was coming out of their chimney. Every after- 
 noon the butcher gave the hungry dog a piece of meat. 
 One evening a carriage was stopping in front of my 
 kitchen. I wear a number thirteen collar. Guess what 
 made me sneeze. Send me a pair of leather shoes. I 
 do not know, but I am almost sure they are mine. My 
 uncle bought my cousin a pretty watch for forty dollars. 
 The soldier dropped his sword. Jack had a whistle and 
 a£so £we£ve nails. The ocean does not often freeze. You 
 should speak to people whom you meet. It takes or% 
 a minute to pass through the gate and across the roaa\ 
 Did you ever hear a /airy laugh? The Awencaw 
 Indian had a saucer without a cup. Neither a pear wor 
 a peach was at the grocery store to-day. Cut up a w/ioZe 
 omori with a handful of beans. My pmwo ?essow was 
 easy. The animal ran info the road and straight against 
 a tree. Give me another sentence which has the word 
 "title" in it. I believe true friends like to be together 
 instead of apart.
 
 114 
 
 Scientific Measurement 
 
 These 100 selected words (printed in italics) were again 
 put into sentences as shown (page 113) and were dictated 
 later to five schools. Great care was taken to insure uni- 
 formity in the administration of the tests. Later 18 addi- 
 tional words were added, making a total of 118 words 
 dictated. The extent to which each of these 118 words was 
 spelled correctly in each grade in each school was deter- 
 mined. Using the data so collected, it was possible to 
 select words which show a regular increase in difficulty, as 
 we pass down from grade to grade. From these words 
 two lists were then selected, each containing 25 words; 
 these are referred to as the "first preferred list" and 
 "second preferred list," as tabulated below. 
 
 PREFERRED LIST 
 
 
 First 
 
 
 
 Secon 
 
 D 
 
 
 1. 
 
 even 
 
 14. 
 
 minute 
 
 26. 
 
 already 
 
 39. 
 
 too 
 
 2. 
 
 lesson 
 
 15. 
 
 cousin 
 
 27. 
 
 beginning 
 
 40. 
 
 towel 
 
 3. 
 
 only 
 
 16. 
 
 nails 
 
 28. 
 
 chicken 
 
 41. 
 
 Tuesday 
 
 4. 
 
 smoke 
 
 17. 
 
 janitor 
 
 29. 
 
 choose 
 
 42. 
 
 tying 
 
 5. 
 
 front 
 
 18. 
 
 saucer 
 
 30. 
 
 circus 
 
 43. 
 
 whole 
 
 6. 
 
 sure 
 
 19. 
 
 stopping 
 
 31. 
 
 grease 
 
 44. 
 
 against 
 
 7. 
 
 pear 
 
 20. 
 
 sword 
 
 32. 
 
 pigeons 
 
 45. 
 
 answer 
 
 8. 
 
 bought 
 
 21. 
 
 freeze 
 
 33. 
 
 quarrel 
 
 46. 
 
 butcher 
 
 9. 
 
 another 
 
 22. 
 
 touch 
 
 34. 
 
 saucy 
 
 47. 
 
 guess 
 
 10. 
 
 forty 
 
 23. 
 
 whistle 
 
 35. 
 
 tailor 
 
 48. 
 
 instead 
 
 11. 
 
 pretty 
 
 24. 
 
 carriage 
 
 36. 
 
 telegram 
 
 49. 
 
 raise 
 
 12. 
 
 wear 
 
 25. 
 
 nor 
 
 37. 
 
 telephone 
 
 50. 
 
 beautiful 
 
 13. 
 
 button 
 
 
 
 38. 
 
 tobacco 
 
 
 
 Considering these 50 words alone, Table II shows the 
 percentage of children from the third to the eighth grade, 
 who were able to spell each of the 50 words. Thus, even 
 was spelled correctly by 59% of children in the third 
 grade, 93% in the sixth, and 97% in the eighth grade.
 
 Spelling Scales 
 
 115 
 
 TABLE II 
 (Showing Standard Scores in Spelling) 
 
 Words 
 
 3d Yr. 
 
 4th Yr. 
 
 5th Yr. 
 
 6th Yr. 
 
 7th Yr. 
 
 8th Yr. 
 
 1. even . . . 
 
 59% 
 
 79% 
 
 89% 
 
 93% 
 
 93% 
 
 97% 
 
 2. lesson . . 
 
 
 37 
 
 72 
 
 83 
 
 91 
 
 94 
 
 96 
 
 3. only . . 
 
 
 65 
 
 75 
 
 89 
 
 95 
 
 97 
 
 99 
 
 4. smoke . . 
 
 
 46 
 
 69 
 
 85 
 
 94 
 
 96 
 
 99 
 
 5. front . . 
 
 
 51 
 
 72 
 
 80 
 
 90 
 
 94 
 
 97 
 
 6. sure . . 
 
 
 47 
 
 55 
 
 69 
 
 78 
 
 89 
 
 94 
 
 7. pear . . 
 
 
 31 
 
 42 
 
 58 
 
 72 
 
 81 
 
 94 
 
 8. bought 
 
 
 40 
 
 65 
 
 79 
 
 91 
 
 94 
 
 97 
 
 9. another . 
 
 
 36 
 
 43 
 
 78 
 
 86 
 
 94 
 
 96 
 
 10. forty . . 
 
 
 49 
 
 62 
 
 65 
 
 72 
 
 83 
 
 87 
 
 11. pretty . . 
 
 
 45 
 
 67 
 
 76 
 
 90 
 
 90 
 
 94 
 
 12. wear . . 
 
 
 35 
 
 49 
 
 61 
 
 74 
 
 84 
 
 93 
 
 13. button 
 
 
 32 
 
 52 
 
 61 
 
 73 
 
 74 
 
 87 
 
 14. minute 
 
 
 26 
 
 38 
 
 62 
 
 77 
 
 86 
 
 92 
 
 15. cousin . . 
 
 
 19 
 
 47 
 
 69 
 
 89 
 
 89 
 
 95 
 
 16. nails . . 
 
 
 43 
 
 58 
 
 71 
 
 87 
 
 92 
 
 96 
 
 17. janitor 
 
 
 19 
 
 42 
 
 58 
 
 81 
 
 81 
 
 90 
 
 18. saucer . . 
 
 
 11 
 
 29 
 
 42 
 
 58 
 
 79 
 
 81 
 
 19. stopping . 
 
 
 27 
 
 39 
 
 55 
 
 71 
 
 76 
 
 84 
 
 20. sword . . 
 
 
 13 
 
 46 
 
 57 
 
 78 
 
 86 
 
 93 
 
 21. freeze . . 
 
 
 29 
 
 46 
 
 68 
 
 83 
 
 86 
 
 94 
 
 22. touch . . 
 
 
 45 
 
 52 
 
 60 
 
 81 
 
 84 
 
 93 
 
 23. whistle . 
 
 
 22 
 
 55 
 
 56 
 
 64 
 
 75 
 
 85 
 
 24. carriage . 
 
 
 13 
 
 40 
 
 50 
 
 67 
 
 81 
 
 85 
 
 25. nor . . . 
 
 
 63 
 
 61 
 
 65 
 
 68 
 
 77 
 
 94 
 
 26. already 
 
 
 16 
 
 42 
 
 43 
 
 62 
 
 44 
 
 77 
 
 27. beginning 
 
 
 9 
 
 25 
 
 37 
 
 46 
 
 66 
 
 75 
 
 28. chicken 
 
 
 49 
 
 70 
 
 83 
 
 90 
 
 96 
 
 99 
 
 29. choose 
 
 
 22 
 
 34 
 
 48 
 
 60 
 
 65 
 
 82 
 
 30. circus . . 
 
 
 20 
 
 39 
 
 50 
 
 72 
 
 75 
 
 95 
 
 31. grease . . 
 
 
 11 
 
 18 
 
 37 
 
 35 
 
 42 
 
 57 
 
 32. pigeons 
 
 
 7 
 
 29 
 
 41 
 
 57 
 
 70 
 
 82 
 
 33. quarrel 
 
 
 15 
 
 39 
 
 53 
 
 75 
 
 86 
 
 94 
 
 34. saucy . . 
 
 
 14 
 
 35 
 
 40 
 
 52 
 
 71 
 
 78 
 
 35. tailor . . 
 
 
 38 
 
 55 
 
 70 
 
 75 
 
 81 
 
 84 
 
 36. telegram 
 
 
 15 
 
 31 
 
 39 
 
 63 
 
 73 
 
 84 
 
 37. telephone 
 
 
 8 
 
 35 
 
 48 
 
 67 
 
 83 
 
 87 
 
 38. tobacco 
 
 
 12 
 
 39 
 
 60 
 
 75 
 
 88 
 
 96 
 
 39. too . . 
 
 
 14 
 
 28 
 
 27 
 
 24 
 
 30 
 
 43 
 
 40. towel . 
 
 
 24 
 
 44 
 
 64 
 
 73 
 
 78 
 
 94 
 
 41. Tuesday 
 
 
 46 
 
 70 
 
 67 
 
 80 
 
 87 
 
 91 
 
 42. tying . 
 
 
 44 
 
 58 
 
 70 
 
 68 
 
 76 
 
 87 
 
 43. whole . 
 
 
 17 
 
 43 
 
 64 
 
 78 
 
 84 
 
 90 
 
 44. against 
 
 
 19 
 
 30 
 
 54 
 
 75 
 
 84 
 
 94 
 
 45. answer 
 
 
 27 
 
 47 
 
 67 
 
 86 
 
 90 
 
 97 
 
 46. butcher 
 
 
 33 
 
 59 
 
 69 
 
 85 
 
 90 
 
 97 
 
 47. guess . 
 
 
 20 
 
 32 
 
 49 
 
 67 
 
 77 
 
 85 
 
 48. instead 
 
 
 32 
 
 48 
 
 62 
 
 86 
 
 87 
 
 91 
 
 49. raise 
 
 
 21 
 
 54 
 
 67 
 
 84 
 
 93 
 
 94 
 
 50. beautiful 
 
 
 10 
 
 52 
 
 70 
 
 85 
 
 94 
 
 96
 
 116 Scientific Measurement 
 
 In this way, Buckingham has provided a basis of com- 
 parison, which may be used by any teacher, as a method 
 of testing the relative ability of different classes. 1 
 
 DIRECTIONS FOR ADMINISTERING 
 
 The following instructions, which are essentially the 
 same as those followed by Buckingham, may be given as 
 regards the conduct of the test : 
 
 (1) Give all the words in sentences during one session, 
 i.e. either in morning or afternoon of same day, except 
 in classes below the fifth grade, where the material should 
 be given in two periods separated by half an hour at least. 
 
 (2) Each sentence should be dictated, either as a whole 
 or in part, as many times as may seem necessary to secure 
 its complete understanding. This experiment is purely a 
 test in spelling ; it is not expected that the pupils should 
 be subjected to the added difficulty of recalling the words 
 dictated. 
 
 (3) Offer no explanation of separate words or sentences. 
 If the meaning is not clear, repeat the sentence as a whole 
 or in part. 
 
 (4) Do not ask the children to underline words, or 
 otherwise call attention to the significant words of the 
 sentences. 
 
 (5) After the children have written the sentences, read 
 them again, and allow the pupils to insert words or make 
 other corrections before finally collecting the papers. 
 
 These papers may now be collected for the whole class, 
 and the percentage of pupils getting any particular word 
 correct determined and compared with the table which 
 has already been given. Of course no particular signifi- 
 cance is attached to any single word; there is no one 
 word which will test the spelling ability of a group. 
 
 1 The tables in this section are reproduced by the courtesy of Dr. 
 B. R. Buckingham.
 
 Spelling Scales 117 
 
 When, however, 50 words are taken, which have been pre- 
 viously standardized, the manner in which these are 
 spelled by any group of pupils will serve to give a quan- 
 titative idea of their spelling ability. Thus, if it is found 
 by a teacher who is dealing with Grade V, that her aver- 
 age percentage for 50 words falls notably below the aver- 
 age given in the table for Grade V, there is every reason 
 to suppose that there is something abnormal about the 
 standing of that class, due to causes which might profit- 
 ably be investigated. 
 
 Suppose, for example, that we are dealing with a fifth 
 grade which contains 50 children, and we find that the 
 word another is spelled correctly by 31 of the children. 
 Reducing this to the percentage basis, the score of the 
 class for this word is 62%. On reference to Bucking- 
 ham's Table, we see that the average score of this grade 
 for the word another is 78%, which means that the par- 
 ticular grade in question, as far as this word is concerned, 
 was not equal to the average. The same procedure may 
 be repeated with any of the other words in the list, and 
 the average of all the percentages obtained. This figure 
 may then be compared with the averages of the percentage 
 for Grade V given for the particular words employed. It 
 is necessary to use from 10 to 20 words in testing a grade, 
 in order to avoid the danger of picking out one or two 
 words upon which special drill might have been given. 
 When 10 or 20 words are chosen at random from the list, 
 this difficulty is obviated. 
 
 It may appear that some justification is required for 
 this laborious study. The ordinary individual would be 
 apt to take the attitude that the teacher's judgment would 
 be just about as sound as the estimates arrived at by 
 the foregoing process. As a matter of fact, the 50 words 
 were ranked by 300 judges, most of them teachers. 
 Naturally there was a general agreement between the 
 teachers' judgments, and the relative order of the words
 
 118 Scientific Measurement 
 
 found as the result of experimental study. But with 
 certain words, there was very great disagreement. Thus, 
 the word nor when ranked by the teachers was given 
 fifth place as regards ease of spelling. The actual records 
 show that the children found it the sixteenth word as re- 
 gards ease of spelling. Again, the word button was ranked 
 ninth by the teachers, and thirty-first by the records 
 which came from the pupils. This shows the unsatis- 
 factoriness of relying on teachers' judgments. As long 
 as those who are teaching do not know the relative diffi- 
 culty of the words taught, how can they be expected to 
 give the correct weight either in time or emphasis in their 
 teaching? 
 
 Buckingham, in the latter half of his study, proceeds to 
 construct a scale for the measurement of spelling effi- 
 ciency, a scale which contains at one end words which, 
 if they cannot be spelled, would indicate zero ability, 
 and at the other end words which are very difficult for 
 the average child in the grades to spell. By simple statis- 
 tical methods and suitable assumptions he determined 
 the interval between the words on the scale, the length 
 of the interval being measured by the increase in diffi- 
 culty as shown by the percentage of times it was correctly 
 spelled. It would be impossible in the limits of this book 
 to explain the method of derivation of the scale. Its 
 interest is largely theoretical, and in its present form 
 it could not be used with profit by the average teacher. 
 It should, however, be borne in mind that such a measur- 
 ing rod has been constructed even in a difficult function 
 such as spelling.
 
 Spelling Scales 119 
 
 II. STARCH SPELLING SCALE 
 
 A second method of measuring spelling ability has been 
 devised by Starch, who worked quite independently of 
 Buckingham. While this method lacks the statistical 
 precision of Buckingham's study, in that it assumes (as 
 far as the score is concerned) each word to be of equal 
 difficulty, it is very straightforward and has many points 
 to recommend its use in the classroom. The first object 
 of the experiment was to obtain six lists of equal diffi- 
 culty, each containing 100 words, representative of the 
 entire non-scientific English vocabulary. This was ac- 
 complished by taking at random the first defined word of 
 more than two letters on every even-numbered page in 
 Webster's New International Dictionary. This made a 
 total of 1,186 words. Every technical, psychological and 
 obsolete word was then discarded, leaving 600 words. 
 These were then arranged alphabetically in the order of 
 size beginning with three-letter words, four-letter words, 
 etc. This list was then divided into six lists of 100 words 
 each, by choosing for the first list, the first, seventh, 
 thirteenth, etc., word of the original list of 600 words. 
 The second list was obtained in a similar manner by tak- 
 ing the second, eighth, and fourteenth word, etc. ; and 
 so on till the sixth list, which was formed by taking the 
 sixth and twelfth word, and so on. The lists which re- 
 sulted from this process are as follows :
 
 120 
 
 Scientific Measurement 
 
 
 
 
 LIST I 
 
 
 
 1. 
 
 add 
 
 35. 
 
 prism 
 
 69. 
 
 commence 
 
 2. 
 
 but 
 
 36. 
 
 rogue 
 
 70. 
 
 estimate 
 
 3. 
 
 get 
 
 37. 
 
 shape 
 
 71. 
 
 flourish 
 
 4. 
 
 low 
 
 38. 
 
 steal 
 
 72. 
 
 luckless 
 
 5. 
 
 rat 
 
 39. 
 
 swain 
 
 73. 
 
 national 
 
 6. 
 
 sun 
 
 40. 
 
 title 
 
 74. 
 
 pinnacle 
 
 7. 
 
 alum 
 
 41. 
 
 wheat 
 
 75. 
 
 reducent 
 
 8. 
 
 blow 
 
 42. 
 
 accrue 
 
 76. 
 
 standing 
 
 9. 
 
 cart 
 
 43. 
 
 bottom 
 
 77. 
 
 venturer 
 
 10. 
 
 cone 
 
 44. 
 
 chapel 
 
 78. 
 
 ascension 
 
 11. 
 
 easy 
 
 45. 
 
 dragon 
 
 79. 
 
 dishallow 
 
 12. 
 
 fell 
 
 46. 
 
 filter 
 
 80. 
 
 imposture 
 
 13. 
 
 foul 
 
 47. 
 
 hearse 
 
 81. 
 
 invective 
 
 14. 
 
 gold 
 
 48. 
 
 laden 
 
 82. 
 
 rebellion 
 
 15. 
 
 head 
 
 49. 
 
 milden 
 
 83. 
 
 scrimping 
 
 16. 
 
 kiss 
 
 50. 
 
 pilfer 
 
 84. 
 
 unalloyed 
 
 17. 
 
 long 
 
 51. 
 
 rabbit 
 
 85. 
 
 volunteer 
 
 18. 
 
 mock 
 
 52. 
 
 school 
 
 86. 
 
 cardinally 
 
 19. 
 
 neck 
 
 53. 
 
 shroud 
 
 87. 
 
 connective 
 
 20. 
 
 rest 
 
 54. 
 
 starch 
 
 88. 
 
 effrontery 
 
 21. 
 
 spur 
 
 55. 
 
 vanity 
 
 89. 
 
 indistinct 
 
 22. 
 
 then 
 
 56. 
 
 bizarre 
 
 90. 
 
 nunciature 
 
 23. 
 
 vile 
 
 57. 
 
 compose 
 
 91. 
 
 sphericity 
 
 24. 
 
 afoot 
 
 58. 
 
 dismiss 
 
 92. 
 
 attenuation 
 
 25. 
 
 black 
 
 59. 
 
 faction 
 
 93. 
 
 fulminating 
 
 26. 
 
 brush 
 
 60. 
 
 hemlock 
 
 94. 
 
 lamentation 
 
 27. 
 
 close 
 
 61. 
 
 leopard 
 
 95. 
 
 secretarial 
 
 28. 
 
 dodge 
 
 62. 
 
 omnibus 
 
 96. 
 
 apparitional 
 
 29. 
 
 faint 
 
 63. 
 
 procure 
 
 97. 
 
 intermissive 
 
 30. 
 
 force 
 
 64. 
 
 rinsing 
 
 98. 
 
 subjectively 
 
 31. 
 
 grape 
 
 65. 
 
 splashy 
 
 99. 
 
 inspirational 
 
 32. 
 
 honor 
 
 66. 
 
 torpedo 
 
 100. 
 
 ineffectuality 
 
 33. 
 
 mince 
 
 67. 
 
 worship 
 
 
 
 34. 
 
 paint 
 
 68. 
 
 bescreen 
 

 
 Spelling Scales 
 
 121 
 
 1. air 
 
 2. cat 
 
 3. hop 
 
 4. man 
 
 5. row 
 
 6. tap 
 
 7. awry 
 
 8. blue 
 
 9. cast 
 
 10. corn 
 
 11. envy 
 
 12. feud 
 
 13. game 
 
 14. grow 
 
 15. home 
 
 16. knee 
 
 17. look 
 
 18. mold 
 
 19. part 
 
 20. ruin 
 
 21. take 
 
 22. tree 
 
 23. well 
 
 24. allay 
 
 25. blaze 
 
 26. buggy 
 
 27. clown 
 
 28. doubt 
 
 29. false 
 
 30. forth 
 
 31. grass 
 
 32. house 
 
 33. money 
 
 34. paper 
 
 LIST II 
 
 35. quill 
 
 36. rough 
 
 37. shout 
 
 38. stick 
 
 39. swear 
 
 40. trump 
 
 41. whirl 
 
 42. action 
 
 43. bridle 
 
 44. charge 
 
 45. driver 
 
 46. finger 
 
 47. heaven 
 
 48. legend 
 
 49. motley 
 
 50. portal 
 
 51. recipe 
 
 52. scrape 
 
 53. simple 
 
 54. strain 
 
 55. weaken 
 
 56. breaker 
 
 57. congeal 
 
 58. disturb 
 
 59. foreign 
 
 60. hoggery 
 
 61. meaning 
 
 62. onerate 
 
 63. provoke 
 
 64. salient 
 
 65. station 
 
 66. trample 
 
 67. abstract 
 
 68. bulletin 
 
 69. covenant 
 
 70. eugenics 
 
 71. friskful 
 
 72. luminous 
 
 73. opulence 
 
 74. planchet 
 
 75. reformer 
 
 76. thorough 
 
 77. watering 
 
 78. belonging 
 
 79. displayed 
 
 80. indention 
 
 81. mercenary 
 
 82. redevelop 
 
 83. senescent 
 
 84. uncharged 
 
 85. whichever 
 
 86. centennial 
 
 87. constitute 
 
 88. exaltation 
 
 89. in vocative 
 
 90. personable 
 
 91. strawberry 
 
 92. concentrate 
 
 93. imaginative 
 
 94. mathematics 
 
 95. selfishness 
 
 96. collectivity 
 
 97. marriageable 
 
 98. agriculturist 
 
 99. quarantinable 
 100. relinquishment
 
 122 
 
 Scientific Measurement 
 
 
 
 
 LIST III 
 
 
 
 1. 
 
 art 
 
 35. 
 
 razor 
 
 69. 
 
 dominate 
 
 2. 
 
 dry 
 
 36. 
 
 saint 
 
 70. 
 
 exchange 
 
 3. 
 
 ice 
 
 37. 
 
 smell 
 
 71. 
 
 governor 
 
 4. 
 
 mix 
 
 38. 
 
 stock 
 
 72. 
 
 manifest 
 
 5. 
 
 run 
 
 39. 
 
 swoop 
 
 73. 
 
 osculate 
 
 6. 
 
 top 
 
 40. 
 
 twine 
 
 74. 
 
 pleasure 
 
 7. 
 
 back 
 
 41. 
 
 white 
 
 75. 
 
 revising 
 
 8. 
 
 bond 
 
 42. 
 
 barrel 
 
 76. 
 
 traverse 
 
 9. 
 
 chip 
 
 43. 
 
 buckle 
 
 77. 
 
 westward 
 
 10. 
 
 crib 
 
 44. 
 
 cotton 
 
 78. 
 
 capitally 
 
 11. 
 
 ever 
 
 45. 
 
 engine 
 
 79. 
 
 extremism 
 
 12. 
 
 fire 
 
 46. 
 
 flimsy 
 
 80. 
 
 indicated 
 
 13. 
 
 gilt 
 
 47. 
 
 helmet 
 
 81. 
 
 monoplane 
 
 14. 
 
 hack 
 
 48. 
 
 lesser 
 
 82. 
 
 repertory 
 
 15. 
 
 hunt 
 
 49. 
 
 ocular 
 
 83. 
 
 stimulate 
 
 16. 
 
 lace 
 
 50. 
 
 potato 
 
 84. 
 
 unlocated 
 
 17. 
 
 main 
 
 51. 
 
 relate 
 
 85. 
 
 accidental 
 
 18. 
 
 more 
 
 52. 
 
 season 
 
 86. 
 
 citizenize 
 
 19. 
 
 pelt 
 
 53. 
 
 single 
 
 87. 
 
 contribute 
 
 20. 
 
 sand 
 
 54. 
 
 supply 
 
 88. 
 
 expertness 
 
 21. 
 
 tang 
 
 55. 
 
 weight 
 
 89. 
 
 locomotive 
 
 22. 
 
 turn 
 
 56. 
 
 captain 
 
 90. 
 
 prevailing 
 
 23. 
 
 wine 
 
 57. 
 
 contour 
 
 91. 
 
 symmetrize 
 
 24. 
 
 amuse 
 
 58. 
 
 earnest 
 
 92. 
 
 consolatory 
 
 25. 
 
 blind 
 
 59. 
 
 fowling 
 
 93. 
 
 incremental 
 
 26. 
 
 catch 
 
 60. 
 
 inflate 
 
 94. 
 
 penetrative 
 
 27. 
 
 count 
 
 61. 
 
 measure 
 
 95. 
 
 superintend 
 
 28. 
 
 dress 
 
 62. 
 
 palaver 
 
 96. 
 
 conterminous 
 
 29. 
 
 fancy 
 
 63. 
 
 raising 
 
 97. 
 
 naturalistic 
 
 30. 
 
 freak 
 
 64. 
 
 seizing 
 
 98. 
 
 artificiality 
 
 31. 
 
 gross 
 
 65. 
 
 sulphur 
 
 99. 
 
 re-examination 
 
 32. 
 
 inlet 
 
 66. 
 
 trestle 
 
 100. 
 
 sentimentalism 
 
 33. 
 
 muddy 
 
 67. 
 
 adhesive 
 
 
 
 34. 
 
 peace 
 
 68. 
 
 buttress 
 

 
 Spelling Scales 
 
 123 
 
 
 
 
 LIST IV 
 
 
 
 1. 
 
 bee 
 
 35. 
 
 remit 
 
 69. 
 
 enabling 
 
 2. 
 
 elk 
 
 36. 
 
 scale 
 
 70. 
 
 external 
 
 3. 
 
 key 
 
 37. 
 
 speak 
 
 71. 
 
 greeting 
 
 4. 
 
 new 
 
 38. 
 
 stone 
 
 72. 
 
 mosquito 
 
 5. 
 
 saw 
 
 39. 
 
 thick 
 
 73. 
 
 outfling 
 
 6. 
 
 war 
 
 40. 
 
 under 
 
 74. 
 
 positive 
 
 7. 
 
 base 
 
 41. 
 
 widen 
 
 75. 
 
 romantic 
 
 8. 
 
 book 
 
 42. 
 
 bearer 
 
 76. 
 
 undulate 
 
 9. 
 
 clue 
 
 43. 
 
 canine 
 
 77. 
 
 adverbial 
 
 10. 
 
 down 
 
 44. 
 
 create 
 
 78. 
 
 carpentry 
 
 11. 
 
 fall 
 
 45. 
 
 eraser 
 
 79. 
 
 franchise 
 
 12. 
 
 flat 
 
 46. 
 
 garret 
 
 80. 
 
 infatuate 
 
 13. 
 
 girt 
 
 47. 
 
 hollow 
 
 81. 
 
 promenade 
 
 14. 
 
 hand 
 
 48. 
 
 little 
 
 82. 
 
 rigmarole 
 
 15. 
 
 iron 
 
 49. 
 
 office 
 
 83. 
 
 stripling 
 
 16. 
 
 lime 
 
 50. 
 
 prince 
 
 84. 
 
 vegetable 
 
 17. 
 
 make 
 
 51. 
 
 retain 
 
 85. 
 
 assignment 
 
 18. 
 
 move 
 
 52. 
 
 settle 
 
 86. 
 
 comparison 
 
 19. 
 
 plug 
 
 53. 
 
 sluice 
 
 87. 
 
 coordinate 
 
 20. 
 
 shop 
 
 54. 
 
 swerve 
 
 88. 
 
 expressage 
 
 21. 
 
 tear 
 
 55. 
 
 withal 
 
 89. 
 
 mayonnaise 
 
 22. 
 
 tusk 
 
 56. 
 
 chicken 
 
 90. 
 
 recompense 
 
 23. 
 
 wire 
 
 57. 
 
 counter 
 
 91. 
 
 untraveled 
 
 24. 
 
 apple 
 
 58. 
 
 emperor 
 
 92. 
 
 consumptive 
 
 25. 
 
 blood 
 
 59. 
 
 freight 
 
 93. 
 
 infuriation 
 
 26. 
 
 chain 
 
 60. 
 
 journal 
 
 94. 
 
 photosphere 
 
 27. 
 
 craft 
 
 61. 
 
 neglect 
 
 95. 
 
 terrestrial 
 
 28. 
 
 drawn 
 
 62. 
 
 passion 
 
 96. 
 
 horsemanship 
 
 29. 
 
 field 
 
 63. 
 
 reserve 
 
 97. 
 
 regenerative 
 
 30. 
 
 frost 
 
 64. 
 
 serpent 
 
 98. 
 
 circumscribed 
 
 31. 
 
 guard 
 
 65. 
 
 surface 
 
 99. 
 
 sculpturesque 
 
 32. 
 
 jelly 
 
 66. 
 
 trouble 
 
 100. 
 
 verisimilitude 
 
 33. 
 
 ocean 
 
 67. 
 
 affected 
 
 
 
 34. 
 
 pitch 
 
 68. 
 
 calendar 
 

 
 124 
 
 Scientific Measurement 
 
 
 
 
 LIST V 
 
 
 
 1. 
 
 bow 
 
 35. 
 
 revel 
 
 69. 
 
 entirely 
 
 2. 
 
 fly 
 
 36. 
 
 scorn 
 
 70. 
 
 farewell 
 
 3. 
 
 law 
 
 37. 
 
 spire 
 
 71. 
 
 incident 
 
 4. 
 
 old 
 
 38. 
 
 strut 
 
 72. 
 
 mountain 
 
 5. 
 
 see 
 
 39. 
 
 three 
 
 73. 
 
 parallel 
 
 6. 
 
 ache 
 
 40. 
 
 voice 
 
 74. 
 
 prelimit 
 
 7. 
 
 bead 
 
 41. 
 
 wince 
 
 75. 
 
 spectral 
 
 8. 
 
 call 
 
 42. 
 
 beaver 
 
 76. 
 
 urbanize 
 
 9. 
 
 cold 
 
 43. 
 
 cannon 
 
 77. 
 
 aggrieved 
 
 10. 
 
 draw 
 
 44. 
 
 crispy 
 
 78. 
 
 clarifier 
 
 11. 
 
 fast 
 
 45. 
 
 escape 
 
 79. 
 
 hydraulic 
 
 12. 
 
 foil 
 
 46. 
 
 gladly 
 
 80. 
 
 inheritor 
 
 13. 
 
 glue 
 
 47. 
 
 hustle 
 
 81. 
 
 purgation 
 
 14. 
 
 hard 
 
 48. 
 
 mallet 
 
 82. 
 
 sacrifice 
 
 15. 
 
 jack 
 
 49. 
 
 oriole 
 
 83. 
 
 surviving 
 
 16. 
 
 line 
 
 50. 
 
 pulley 
 
 84. 
 
 vestibule 
 
 17. 
 
 mark 
 
 51. 
 
 rubric 
 
 85. 
 
 authorship 
 
 18. 
 
 musk 
 
 52. 
 
 shears 
 
 86. 
 
 concoction 
 
 19. 
 
 prig 
 
 53. 
 
 solace 
 
 87. 
 
 derigation 
 
 20. 
 
 slat 
 
 54. 
 
 trifle 
 
 88. 
 
 federative 
 
 21. 
 
 test 
 
 55. 
 
 yellow 
 
 89. 
 
 memorandum 
 
 22. 
 
 vend 
 
 56. 
 
 circuit 
 
 90. 
 
 regularity 
 
 23. 
 
 wood 
 
 57. 
 
 crooked 
 
 91. 
 
 abnormality 
 
 24. 
 
 armor 
 
 58. 
 
 enstamp 
 
 92. 
 
 disseminate 
 
 25. 
 
 boast 
 
 59. 
 
 general 
 
 93. 
 
 insensitive 
 
 26. 
 
 chase 
 
 60. 
 
 lateral 
 
 94. 
 
 predominate 
 
 27. 
 
 cross 
 
 61. 
 
 nourish 
 
 95. 
 
 unprevented 
 
 28. 
 
 enjoy 
 
 62. 
 
 placard 
 
 96. 
 
 inarticulate 
 
 29. 
 
 fixed 
 
 63. 
 
 resolve 
 
 97. 
 
 stupendously 
 
 30. 
 
 glean 
 
 64. 
 
 signify 
 
 98. 
 
 communicating 
 
 31. 
 
 guild 
 
 65. 
 
 tabloid 
 
 99. 
 
 anthropometric 
 
 32. 
 
 joint 
 
 66. 
 
 unitive 
 
 100. 
 
 emancipationist 
 
 33. 
 
 order 
 
 67. 
 
 approved 
 
 
 
 34. 
 
 point 
 
 68. 
 
 cerebral 
 

 
 Spelling Scales 
 
 125 
 
 
 
 
 LIST VI 
 
 
 
 1. 
 
 box 
 
 35. 
 
 river 
 
 69. 
 
 erosible 
 
 2. 
 
 gap 
 
 36. 
 
 shaft 
 
 70. 
 
 fetching 
 
 3. 
 
 lay 
 
 37. 
 
 stall 
 
 71. 
 
 juncture 
 
 4. 
 
 pod 
 
 38. 
 
 sugar 
 
 72. 
 
 narcotic 
 
 5. 
 
 sex 
 
 39. 
 
 throw 
 
 73. 
 
 parasite 
 
 6. 
 
 alms 
 
 40. 
 
 watch 
 
 74. 
 
 probator 
 
 7. 
 
 bird 
 
 41. 
 
 young 
 
 75. 
 
 squeaker 
 
 8. 
 
 camp 
 
 42. 
 
 begird 
 
 76. 
 
 vagabond 
 
 9. 
 
 comb 
 
 43. 
 
 causal 
 
 77. 
 
 amphibian 
 
 10. 
 
 dusk 
 
 44. 
 
 discus 
 
 78. 
 
 clearness 
 
 11. 
 
 fear 
 
 45. 
 
 ferret 
 
 79. 
 
 impatient 
 
 12. 
 
 foot 
 
 46. 
 
 gutter 
 
 80. 
 
 intestine 
 
 13. 
 
 goat 
 
 47. 
 
 killed 
 
 81. 
 
 quadruple 
 
 14. 
 
 hawk 
 
 48. 
 
 middle 
 
 82. 
 
 sauciness 
 
 15. 
 
 keep 
 
 49. 
 
 paddle 
 
 83. 
 
 ticketing 
 
 16. 
 
 life 
 
 50. 
 
 puzzle 
 
 84. 
 
 virulence 
 
 17. 
 
 mass 
 
 51. 
 
 sample 
 
 85. 
 
 bafflement 
 
 18. 
 
 navy 
 
 52. 
 
 shield 
 
 86. 
 
 condescend 
 
 19. 
 
 raft 
 
 53. 
 
 spring 
 
 87. 
 
 disconcert 
 
 20. 
 
 some 
 
 54. 
 
 tubule 
 
 88. 
 
 illiterate 
 
 21. 
 
 that 
 
 55. 
 
 bicycle 
 
 89. 
 
 metropolis 
 
 22. 
 
 vice 
 
 56. 
 
 commode 
 
 90. 
 
 repression 
 
 23. 
 
 work 
 
 57. 
 
 discard 
 
 92. 
 
 animalcular 
 
 24. 
 
 aside 
 
 58. 
 
 excuser 
 
 92. 
 
 divestiture 
 
 25. 
 
 brawn 
 
 59. 
 
 gravity 
 
 93. 
 
 intrinsical 
 
 26. 
 
 chime 
 
 60. 
 
 leaping 
 
 94. 
 
 prerogative 
 
 27. 
 
 crown 
 
 61. 
 
 obloquy 
 
 95. 
 
 upholsterer 
 
 28. 
 
 equip 
 
 62. 
 
 pontiff 
 
 96. 
 
 interference 
 
 29. 
 
 flock 
 
 63. 
 
 retreat 
 
 97. 
 
 subantarctic 
 
 30. 
 
 grand 
 
 64. 
 
 society 
 
 98. 
 
 convocational 
 
 31. 
 
 hedge 
 
 65. 
 
 tigress 
 
 99. 
 
 imperturbation 
 
 32. 
 
 knock 
 
 66. 
 
 vitiate 
 
 100. 
 
 irresponsibility 
 
 33. 
 
 ought 
 
 67. 
 
 auditory 
 
 
 
 34. 
 
 poppy 
 
 68. 
 
 churlish 
 
 
 
 These scales 
 
 are reproduced by the courtesy of 
 
 Dr. Daniel Starch.
 
 126 Scientific Measurement 
 
 The advantages of this method of selection are : (1) It 
 gives a random sampling of the entire non-technical Eng- 
 lish vocabulary, for easy words and very hard words 
 occur in the same proportion in the lists as in the English 
 language. (2) The list contains words sufficiently easy 
 to test the poorest speller. (3) The essential requirement 
 of every scientific experiment is fulfilled, since another 
 600 words of the same average difficulty can be chosen, 
 by employing the same method of selection, e.g. the tenth 
 word in the dictionary could be used in place of the first 
 word. 
 
 DIRECTIONS FOR ADMINISTERING TESTS 
 
 First have the pupils write the name, grade, school, 
 city and date at the top of the sheet. 
 
 Pronounce the words clearly, but do not sound them 
 phonetically, or inflect them so as to aid the pupils. Give 
 the meaning of words that sound like words with a dif- 
 ferent meaning and spelling. The pupils are to write the 
 words and to number them in the order in which they are 
 given. Allow sufficient time for the writing. 
 
 Each grade is to be tested twice on two successive days. 
 Use any one of the six lists on the first day and a different 
 list on the second day. (When an entire school is being 
 tested it may be desirable, though not necessary, to use 
 on the first day the same list, say List 1, in all grades, and 
 any other list on the second day.) 
 
 In the first grade use the first 40 words of the list, in 
 the second grade use the first 65 words, in the third grade 
 use the first 80 words, in the fourth grade use the first 90 
 words, and in all other grades use the entire list. 
 
 It has been demonstrated by administering the lists in 
 schools, that each of them is of approximately the same 
 difficulty. It is perhaps desirable, however, when meas- 
 uring the efficiency of an individual group, to give two
 
 JYI 
 
 yiiai 
 
 I2>A 
 
 1M 
 
 
 1 
 
 B/1 
 
 rllfH 
 
 JGAi 
 
 u3m 
 
 
 ) 
 
 
 l 
 
 H 
 
 
 
 1 
 
 3 
 
 J 
 
 4f*i <9* 
 
 /■*^r 
 
 *8 
 
 88 
 
 S€; 
 
 SY 
 
 6T 
 
 3 
 
 88 
 
 se 
 
 ■&e 
 
 ee 
 
 8€ l 
 
 > 
 
 36 
 
 se 
 
 ee 
 
 oor 
 
 -HTHUC 
 30AHL 
 
 
 
 
 
 
 
 
 
 
 
 
 
 > 
 
 •^ - 
 
 ee 
 
 oor 
 
 ,Hnn 
 
 
 5 
 
 
 
 .HTXI8 
 ^30AR3 
 
 
 I 
 
 ir 
 
 
 
 
 I 
 
 k 
 l 
 
 navea 
 
 saic 
 
 ftb 
 
 v> 
 
 \o 
 
 
 Jggio* 
 
 09fil 
 
 tea 
 
 9Vfid 
 
 9d 
 
 
 ^qqfld 
 
 nifli 
 
 Mi 
 
 9TB 
 
 »wd 
 
 
 aoofl 
 
 obh 
 
 Sol 
 
 bad 
 
 •id* 
 
 • 
 
 jlxridJ 
 
 09X* 
 
 nod 
 
 19VO 
 
 Ua 
 
 
 idfeh 
 
 2fo» 
 
 jcoisd 
 
 izum 
 
 TOO^ 
 
 
 12B9 
 
 log 
 
 100b 
 
 oieai 
 
 tuo 
 
 
 biftf 
 
 xllioa 
 
 wol 
 
 loodoE 
 
 9£0ij 
 
 
 riiuoa 
 
 9tidw 
 
 tsaiJa 
 
 ^B£0 
 
 
 qaab 
 
 txwqe 
 
 rtos 
 
 fta 
 
 olixi 
 
 
 5fa:3ai 
 
 iool 
 
 bneta 
 
 9XU03 
 
 mid 
 
 
 ould 
 
 wold 
 
 bis^ 
 
 boBd 
 
 ^Bbo) 
 
 
 laog 
 
 rfDOid 
 
 jnhd 
 
 son 
 
 ifool 
 
 
 awot 
 
 jahqe 
 
 Us) 
 
 9VU 
 
 bib 
 
 
 ^BJ8 
 
 TOvn 
 
 svQ 
 
 llLrf 
 
 e:4ii 
 
 
 bnnia 
 
 laalq 
 
 Did 
 
 ttel 
 
 /■s. 
 
 
 D. ; ?JI)0 
 
 tm 
 
 wai 
 
 to! 
 
 ^od 
 
 
 lixab 
 
 Snog 
 
 As 
 
 aid 
 
 Jlood 
 
 
 ba£d 
 
 I9*niw 
 
 iou'r 
 
 leriiom 
 
 
 
 •mag 
 
 JBOd 
 
 320Ji 
 
 ^aw 
 
 S9ldJ 
 
 
 
 99Tl 
 
 J»8 
 
 baef 
 
 
 
 )8»1 
 
 oJhsI 
 
 oincd 
 
 bloi 
 
 
 
 J2B9 
 
 93fiq 
 
 daum 
 
 tod 
 
 
 
 no3 
 
 ooia 
 
 Wo 
 
 lad 
 
 
 
 ql:»d 
 
 bfl9 
 
 gnol 
 
 blirio 
 
 
 
 bx&d 
 
 DbI 
 
 970I 
 
 931 
 
 
 
 ODBl 
 
 J99* 
 
 09C& 
 
 ^Blq 
 
 
 
 19V09 
 
 ltt9W 
 
 o?.uod 
 
 •98 
 
 
 
 9xd 
 
 Jij,.ij 
 
 T89^ 
 
 
 
 
 93B 
 
 ^BWB 
 
 01 
 
 

 
 MEASURING SCALE FOR ABILITY IN SPELLING 
 
 SECOND 
 GRADE 
 
 A 
 
 B 
 
 c 
 
 D 
 
 E 
 
 F 
 
 G 
 
 H 
 
 1 
 
 J 
 
 K 
 
 L 
 
 M 
 
 N 
 
 o 
 
 P 
 
 Q 
 
 R 
 
 s 
 
 T 
 
 u 
 
 V 
 
 w 
 
 X 
 
 Y 
 
 z 
 
 • 99 
 
 98 
 
 96 
 
 94 
 
 92 
 
 88 
 
 84 
 
 79 
 
 73 
 
 66 
 
 58 
 
 Cj-i -.SECOND 
 
 
 
 .THIRD 
 
 GHAHE 
 
 
 
 -FOURTH 
 
 GRADE 
 
 
 «.FIFTH 
 *GRAD£ 
 
 
 J.6IXTH 
 ^BRADE 
 
 
 ^SEVENTH 
 
 
 
 THIRD^. 
 GRADE 
 
 100 
 
 99 
 
 98 
 
 96 
 
 94 
 
 92 
 
 88 
 
 84 
 
 79 
 
 73 
 
 66 
 
 58 
 
 50 
 
 
 
 FOURTH. 
 6RA0E 
 
 100 
 
 99 
 
 98 
 
 96 
 
 94 
 
 92 
 
 88 
 
 84 
 
 79 
 
 73 
 
 66 
 
 58 
 
 50 
 
 
 FIFTH. 
 GRADE 
 
 100 
 
 99 
 
 98 
 
 96 
 
 94 
 
 92 
 
 88 
 
 84 
 
 79 
 
 73 
 
 66 
 
 58 
 
 50 
 
 
 SIXTH. 
 SHADE 
 
 100 
 
 99 
 
 98 
 
 96 
 
 94 
 
 92 
 
 88 
 
 84 
 
 79 
 
 73 
 
 66 
 
 58 
 
 50 
 
 
 SEVENTH. 
 GRADE 
 
 100 
 
 99 
 
 98 
 
 96 
 
 94 
 
 92 
 
 88 
 
 84 
 
 79 
 
 73 
 
 66 
 
 58 
 
 50 
 
 
 EIGHTH. 
 
 GRADE ' 
 
 100 
 
 99 
 
 98 
 
 96 
 
 94 
 
 92 
 
 88 
 
 84 
 
 79 
 
 73 
 
 66 
 
 58 
 
 50 
 
 
 ■rt 
 
 
 the 
 
 be 
 
 Of 
 
 by 
 
 day 
 
 nine 
 
 Mean. 
 
 became 
 
 catch 
 
 trust 
 
 except 
 
 eight 
 
 speod 
 
 sometimes 
 
 forenoon 
 
 often 
 
 guess 
 
 meant 
 
 principal 
 
 organisation 
 
 Immediate 
 
 decision 
 
 Judgment 
 
 
 
 
 
 will 
 
 
 
 eat 
 
 face 
 
 
 brother 
 
 black 
 
 
 
 
 enjoy 
 
 
 
 stopped 
 
 
 
 testimony 
 
 emergency 
 
 convenient 
 
 
 recommend 
 
 
 
 ■t 
 
 
 
 
 
 all 
 
 
 happy 
 
 
 
 
 
 
 
 eogage 
 
 combination 
 
 argument 
 
 whether 
 
 
 appreciate 
 
 receipt 
 
 
 allege 
 
 
 
 
 
 
 this 
 
 had 
 
 lot 
 
 ride 
 
 
 
 
 
 
 rather 
 
 
 
 
 theater 
 
 volume 
 
 distinguish 
 
 arrangement 
 
 sincerely 
 
 pielimlnary 
 
 
 
 
 
 
 
 
 all 
 
 
 box 
 
 tree 
 
 
 
 do thing 
 
 teach 
 
 
 comfort 
 
 complaint 
 
 terrible 
 
 neighbor 
 
 
 
 
 athletic 
 
 disappoint 
 
 
 
 
 
 
 
 
 
 
 tick 
 
 alstef 
 
 
 began 
 able 
 
 
 
 elect 
 
 
 peZd" 
 
 
 century 
 
 
 colonies 
 
 evidence 
 
 extreme 
 
 especially 
 
 
 
 
 
 
 
 
 
 
 got 
 
 cast 
 
 gift S3 
 
 ^Ct 
 
 
 
 
 
 total 
 
 official 
 
 
 experience 
 
 practical 
 
 
 
 
 
 bed 
 
 but 
 
 
 
 vee 
 
 
 
 gone 
 
 
 jail 
 
 beautiful 
 
 addition 
 
 entertain 
 
 
 victim 
 
 relief 
 
 
 proceed 
 
 committee 
 
 
 
 
 
 
 
 
 white 
 
 south 
 
 party 
 
 suit 
 
 file 
 
 built 
 
 Shed 
 
 flight 
 
 
 
 
 estimate. 
 
 probably 
 
 secretary 
 
 association 
 
 cordially 
 
 
 
 
 
 Into 
 
 —J 
 
 soft 
 
 
 
 
 track 
 
 provide 
 
 
 
 
 property 
 
 
 supply 
 assist 
 
 accident 
 
 character 
 
 
 
 
 
 him 
 
 
 
 foot 
 
 inside 
 
 
 
 sight 
 
 
 
 
 
 publication 
 
 Invitation 
 
 foreign 
 
 height | February 
 
 
 
 
 little 
 
 
 
 brtag 
 
 blow 
 
 
 they 
 
 
 stood 
 
 
 district 
 
 
 connect] i) d 
 
 machine 
 
 difference 
 
 
 
 
 
 
 look 
 
 ring 
 
 block 
 
 
 
 fell 
 
 fix 
 
 
 
 
 firm 
 
 
 examination 
 
 impossible 
 
 responsible 
 
 
 
 
 
 did 
 
 live 
 
 tell 
 
 spring 
 
 
 any 
 
 fight 
 
 
 
 
 
 
 
 particular 
 
 concern 
 
 beginning 
 
 
 
 
 old 
 
 like 
 
 trill 
 
 five 
 
 
 stay 
 
 
 
 goes 
 
 death 
 
 objection 
 
 Importance 
 
 
 
 affair 
 
 associate 
 
 application 
 difficulty 
 
 
 
 
 bad 
 
 
 lata 
 
 ball 
 
 plant 
 
 grand 
 
 ahould 
 
 walk 
 
 hold 
 
 
 pleasure 
 
 
 
 
 
 automobile 
 
 
 
 
 red 
 
 
 let 
 
 law 
 
 
 outajde 
 
 o% 
 
 drill 
 
 
 
 
 command 
 
 
 
 various 
 
 
 
 
 
 book 
 
 big 
 
 aak 
 
 song 
 
 dark 
 
 grant 
 
 
 tire 
 
 
 fortune 
 
 debate 
 
 
 local 
 
 
 finally 
 
 
 
 
 mother 
 
 just 
 
 
 
 
 
 
 
 population 
 
 
 
 
 marriage 
 
 entitle 
 
 
 
 
 
 way 
 
 StuDo 
 
 tsr 
 
 
 
 
 check 
 
 Edge' 
 
 TJ," 
 
 publish 
 
 
 further 
 
 political 
 
 
 
 
 
 load 
 
 rJL 
 
 
 first 
 
 
 Income 
 
 prove 
 
 prepare 
 
 
 national 
 
 
 
 
 
 cold 
 
 take 
 
 rett 
 
 
 
 bought 
 paid 
 
 
 beg 
 
 represent 
 
 
 
 
 material 
 
 
 
 
 hat 
 
 
 page 
 
 east 
 
 
 summer 
 
 
 
 
 
 prefer 
 
 condition 
 
 business 
 
 
 
 
 
 bat 
 
 call 
 
 
 SOD 
 
 
 
 
 itself 
 
 contain 
 
 
 
 government 
 
 
 
 
 
 
 child 
 
 
 cod 
 
 help 
 
 
 
 railroad 
 
 
 
 engine 
 
 
 illustrate 
 
 opinion 
 
 
 
 
 
 
 Ice 
 
 
 fall 
 
 bard 
 
 without 
 
 
 unable 
 
 something 
 
 sudden 
 
 visit 
 
 progress 
 entire 
 
 different 
 
 believe 
 
 ought 
 
 
 
 ) 
 
 
 
 then 
 
 feel 
 
 
 afternoon 
 
 
 ticket 
 
 write 
 
 
 
 object 
 
 
 
 maim 1 fully 
 
 
 
 ■ea 
 
 
 went 
 
 
 Priday 
 
 half 
 
 account 
 
 
 instead 
 
 department 
 
 
 
 
 
 agreement 
 
 
 
 r 
 
 back 
 
 On 
 
 
 father 
 
 
 
 throw 
 
 
 
 
 
 Wednesday 
 
 unfortunate 
 
 
 
 
 
 age 
 
 wife 
 
 anything 
 
 real 
 
 thus 
 
 personal 
 
 family 
 
 famous 
 
 already 
 
 certain 
 
 really 
 
 majority 
 
 
 
 
 l 
 
 paper 
 
 gold 
 
 state 
 
 table 
 
 
 woman 
 
 everything 
 
 
 
 attention 
 
 
 celebration 
 
 
 
 
 
 
 
 ffl 
 
 high 
 
 
 
 
 Mrs. 
 
 
 education 
 
 investigate 
 
 fours 
 
 
 
 
 
 
 each 
 
 floe 
 
 talk 
 
 
 fair 
 
 chief 
 
 
 remember 
 
 director 
 
 therefore 
 
 
 necessary 
 
 
 
 
 
 
 
 
 
 
 dollar 
 
 
 
 either 
 
 purpose 
 
 too 
 
 
 divide 
 
 
 
 
 has 
 
 
 May 
 
 2S 
 
 right 
 
 
 
 
 human 
 
 effort 
 
 
 pleasant 
 
 
 
 
 
 
 
 line 
 
 date 
 
 
 brots 
 
 slide 
 
 
 Important 
 
 diamond 
 
 
 
 
 
 
 if 
 
 
 left 
 
 lady 
 
 
 
 farther 
 
 election 
 
 
 
 
 
 
 
 
 
 Monday 
 
 •hip 
 
 
 March 
 
 contract 
 
 feel 
 
 duty 
 
 derk 
 
 Include 
 
 convention 
 
 
 
 
 
 
 her 
 
 yet 
 
 
 better 
 
 
 deal 
 
 
 
 though 
 
 running 
 
 
 
 
 
 
 
 
 Pad 
 
 
 
 
 
 least 
 
 company 
 
 o'clock 
 
 
 manner 
 
 
 
 
 
 
 
 trft 
 
 & 
 
 
 
 brought 
 
 
 
 support 
 
 position 
 field 
 
 
 
 
 
 
 
 baby 
 well 
 
 
 
 herself 
 
 less 
 
 
 
 
 article 
 
 
 
 
 
 
 letter 
 
 
 price 
 
 
 
 God 
 
 
 
 ledge 
 
 
 
 
 
 
 
 
 take 
 
 
 
 oft 
 
 
 
 
 
 effect 
 
 
 
 
 
 
 
 Mr. 
 
 why 
 
 
 
 
 November 
 
 
 
 primary 
 
 
 
 
 
 
 for 
 
 alter 
 
 MB 
 
 
 
 took 
 
 subject 
 
 April 
 
 appear 
 
 Liberty 
 
 
 result 
 
 distribute 
 
 
 
 
 
 
 
 thing 
 
 
 
 country 
 
 
 length 
 
 Saturday 
 
 
 
 
 
 
 Ail the words in each column are of approximately equal spelling 
 
 
 what 
 
 girl 
 
 try 
 
 
 inform 
 
 
 enough 
 
 destroy 
 
 appoint 
 
 information 
 
 
 
 
 
 
 difficulty. The steps in spelling difficulty from each column to the 
 next are approximately equal steps. The numbers at the top indicate 
 
 that 
 bis 
 led 
 
 It* 
 
 toll 
 
 delay 
 behind 
 
 another 
 list 
 
 both 
 heart 
 
 mouth 
 
 himself 
 
 fact 
 September 
 
 newspaper 
 daughter 
 
 consider 
 complete 
 
 
 
 
 
 about what pei cent of correct spellings may be expected among the 
 
 lay 
 
 
 
 people 
 
 children 
 
 
 
 reply 
 
 themselves 
 
 
 
 
 
 
 children of the different grades. For example, if 20 words from 
 
 
 
 
 around 
 
 
 build 
 
 
 attend 
 
 
 
 
 
 
 
 
 column H are given as a spelling test it may be expected that the 
 average score for an entire second grade spelling them will be about 
 
 
 sold 
 
 aide 
 
 kind 
 
 «»p 
 
 held 
 
 understand 
 
 thought 
 
 between 
 
 dUcs 
 
 SEJ" 
 
 popular 
 Christmas 
 
 
 
 Russell Sage Foundation, New York City 
 
 
 79 per cent For a third grade it should be about 92 per cent, for a 
 
 
 told 
 
 life 
 
 dear 
 
 own 
 
 January 
 
 during 
 
 several 
 
 justice 
 
 
 
 
 Division of Education 
 
 
 fourth grade about 98 per cent, and for a fifth grade about 100 per 
 
 
 beat 
 
 
 
 
 member 
 
 
 through 
 
 desire 
 
 gentleman 
 
 
 
 
 
 cent. 
 
 The limits of the groups are as follows: SO means from 46 
 through 54 per cent) 58 means from 55 through 62 per cent; 66 
 means from 63 through 69 per cent; 73 means from 70 through 76 
 percent; 79 means from 77 through 81 per cent; 84 means from 82 
 through 86 per cent; 88 means from 87 through 90 per cent; 92 
 means from 91 through 93 per cent; 94 means 94 and 95 per cent; 
 96 means 96 and 97 per cent; while 98,99 and 100 per cent are sepa- 
 rate groups. 
 
 By means of these groupings a child's spelling ability may be 
 located in terms of grades. Thus if a child were given a 20 word 
 spelling test from the words of column and spelled 15 words, or 75 
 
 
 far 
 
 alike 
 add 
 
 said 
 
 tonight 
 
 tenth 
 
 sir 
 
 these 
 
 dub 
 
 leave 
 
 ground 
 such 
 
 while 
 
 (hose 
 
 Miss 
 died 
 
 copy 
 
 been 
 yesterday 
 among 
 question 
 
 police 
 madam 
 
 address 
 
 August 
 Tuesday 
 
 
 enclose 
 
 suppose 
 wonderful 
 
 direction 
 
 although 
 attempt 
 statement 
 
 
 
 
 Leonard P. Ayres, Director 
 
 The data of this scale are computed from an aggregate of 1 ,400,000 
 spellings by 70,000 children in 84 cities throughout the country. The 
 words are 1,000 in number and the list is the product of combining 
 different studies with the object of identifying the 1,000 common- 
 est words in English writing. Copies of this scale may be obtained 
 :or five cents apiece. Copies of the monograph describing the inves- 
 
 
 
 
 wind 
 ah 
 
 felt 
 full 
 fail 
 
 however 
 
 picture 
 
 December 
 tax 
 
 gutting 
 
 
 written 
 
 
 
 
 tigations which produced it may be obtained for 30 cents each, 
 including the scale. Address the Russell Sage Foundation. Divi- 
 sion of Education, 130 East 22d Street, New York City. 
 
 
 per cent of them, correctly it would be proper to say that he showed 
 
 
 
 nil 
 
 set 
 
 lhall 
 
 
 number 
 
 
 
 arrango 
 
 
 
 
 
 
 fourth grade spelling ability, If he spelled correctly 17 words, or 
 
 
 
 lost 
 
 light 
 comma 
 
 third 
 
 ready 
 
 October 
 
 
 
 
 
 
 
 
 
 85 per cent, he would show fifth grade ability, and so on. 
 
 
 
 anyway 
 
 filth 
 
 
 
 
 
 
 
 
 
 
 
 
 top? 
 
 night 
 
 Eotat 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 paaa 
 
 within 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 glad 
 
 shut 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 trtth 
 
 •**J 
 
 body 
 
 
 
 
 
 
 
 
 
 
 o
 
 Spelling Scales 127 
 
 of the tests. The average of the score made in the two 
 tests will represent pretty accurately the spelling ability. 
 
 STANDARDS OF EFFICIENCY IN SPELLING 
 
 These spelling tests have been standardized by admin- 
 istering them to 2500 pupils in 12 schools of 5 cities, 
 located in Wisconsin, Minnesota, and New York. The 
 average results obtained are shown in the table below, in 
 which the scores are given in round figures. 
 
 Standard Scores for Spelling 
 Grades . . 
 
 1 
 
 2 
 
 3 
 
 4 
 
 5 
 
 6 
 
 7 
 
 8 
 
 10 
 
 30 
 
 40 
 
 51 
 
 61 
 
 71 
 
 78 
 
 85 
 
 This table shows that on the average in Grade III in 
 the schools measured, 40% of each list was spelled cor- 
 rectly. The point of most importance for the individual 
 teacher is to know how the pupils of a particular grade 
 compare in spelling efficiency with pupils of the same 
 grade of other schools. 
 
 By using this very simple device, a purely objective meas- 
 ure of spelling ability can be obtained by the ordinary 
 teacher. No longer need we speak of "good spellers," 
 "bad spellers" and "medium spellers"; we can assign a 
 numerical value to the spelling ability of each individual. 
 
 III. THE AYRES SPELLING TEST (iooo WORDS) 1 
 
 Ayres has also presented a further method of measur- 
 ing spelling ability based on the one thousand most com- 
 mon words in the English language. These words were 
 chosen by combining the results of four previous investiga- 
 tions which had as their object the selection of the words 
 most commonly used in different sorts of writing. The 
 first study was founded on passages from the Bible and 
 other well-known writings, including in all about 100,000 
 
 1 The Ayres Spelling Scale (see insert) is reproduced by the courtesy 
 of Dr. Leonard P. Ayres.
 
 128 Scientific Measurement 
 
 words. The second study of the frequency of different 
 words was made on the basis of an analysis of the words 
 used in 250 different articles taken from issues of four 
 Sunday newspapers published in Buffalo. These articles, 
 counting repetitions, contained 43,989 words; without 
 repetitions, 6000 words. The third study consisted of 
 the tabulation of 23,629 words from 2000 short letters 
 written by 2000 people. The last study comprised a 
 tabulation of some 200,000 words taken from the family 
 correspondence of thirteen adults. 
 
 The list of 1000 words finally selected was determined 
 by combining the results of all these studies. Thus, the 
 1000 words chosen were those which occurred most fre- 
 quently in passages selected from a wide variety of 
 sources; namely, the Bible, the writings of famous 
 authors, newspaper articles, and private correspondence. 
 
 The method employed in standardizing the difficulty 
 of each of the 1000 words was essentially the same as that 
 used by Buckingham, but on a more extensive scale. 
 The 1000 words were first made into 50 lists of 20 words 
 each, and these lists were then administered, in the middle 
 of the school year, to various grades in the schools of 84 
 cities scattered throughout the United States. The data 
 secured from these tests made an aggregate of 1,400,000 
 spellings by 70,000 children. It was on the basis of this 
 data that the Ayres Scale was constructed. 
 
 The scale presented explains itself. All the words in 
 any particular column are of approximately the same 
 spelling difficulty, the difficulty of each word having been 
 determined by the percentage of times the word was 
 spelled correctly in the tests mentioned above. 
 
 DIRECTIONS FOR ADMINISTERING 
 
 The details for administering the tests will be clear from 
 the following example. Suppose we wished to measure
 
 Spelling Scales 129 
 
 the spelling ability of any fifth grade. Taking any one 
 of the columns given in the scale — say Column — we 
 would first of all select any twenty words from it. Then 
 we would dictate these words in a list to the class, giving 
 ample time for each word and explaining the meaning of 
 a word, if doubtful, by putting it in a sentence. Lastly, 
 we would collect the papers and calculate the number of 
 words spelled correctly. If there were 30 children in the 
 class, that would mean that 600 spellings were performed. 
 Suppose out of these 600 spellings there were 480 correct. 
 Then 80% of the words would be correctly spelled. A 
 reference to the scale, Column O, shows that the fifth 
 grade average at midyear is 84%, and the fourth grade 
 average, 73%. Therefore the class measured would be 
 a little below the average fifth grade standing. Suppose 
 a particular child in the grade gets 18 correct out of the 
 20 words. This means a score of 90%, or slightly below 
 the average for the sixth grade, which is 92%. The only 
 care that must be taken in administering the test is not 
 to select a list of words so short that there is a chance of 
 not obtaining representative results. For this reason, in 
 testing the ability of a particular pupil it is well not to 
 use less than 20 words ; but if a group is being tested, so 
 as to obtain merely the group average, a smaller number 
 of words may be used. 
 
 It should be noted that the standards published with 
 the Ayres Scale only apply where these words have been 
 given to pupils who have had no especial drill on them. 
 For, since the words in the scale are so common that 
 they form an excellent foundation for spelling, it is 
 reasonable to suppose that special attention will be 
 given them. This drill will make the pupil too familiar 
 with them to have his score judged by the standard 
 score as obtained by Ayres. This means that probably 
 it will be necessary for each school to establish its own 
 standards.
 
 130 Scientific Measurement 
 
 EXERCISES 
 
 1. Select 15 words from the Buckingham Scale and use these for 
 measuring the spelling ability of a particular class. Outline the steps 
 you would take, and the way in which you would administer the test, 
 score the papers, and tabulate the results. 
 
 2. What are the advantages derived from knowing the relative 
 difficulties of different words? How should this alter the method of 
 teaching? 
 
 3. Using the Starch Scale, how would you establish norms for the 
 grades of your own school? Is it fair to expect a foreign district 
 school and an English-speaking district school to produce the same 
 percentages ? 
 
 4. Suppose a teacher took any list of 100 words and administered 
 these to aggrade and discovered that on the average 75% of the spell- 
 ings were correct, what would this tell or fail to tell the teacher? 
 
 5. If it was found that the average scores of a grade V, for suc- 
 cessive years, tested in January on the Starch Scale, were 59, 60, 61, 
 62, 60, and the average fell suddenly to 53, where would you look for 
 the cause? 
 
 6. How, by means of these scales, would it be possible to compare 
 two different methods of teaching spelling? 
 
 7. If it is found that some children are very much better than the 
 average for their grades, how should this affect the amount of time 
 they devote to spelling? What should be done for those who are 
 much poorer than the average? 
 
 8. Use (a) the Buckingham Scale, (6) the Ayres Scale, (c) the 
 Starch Scale, to test the same class on successive days. Do the re- 
 sults agree, in that they show that the class has the same ability, 
 measured by the grade norms? 
 
 9. Why would it not be fair to apply any of these tests if the 
 children had been drilled on the lists used in these tests? Which is 
 the safest scale to use if we wish to eliminate this error? 
 
 10. Administer List 1 and List 2 of the Starch Scale to the same 
 class, on successive days, and compare the average scores in each. 
 Should they be the same? Why?
 
 CHAPTER VI 
 COMPOSITION SCALES 
 
 I. HILLEGAS SCALE 
 H. HARVARD-NEWTON SCALES 
 
 The task of evaluating efficiency in composition is 
 obviously a complex one because, not only are there 
 several distinct types of composition, such as narration, 
 description, etc., but merit in each of these types is the 
 resultant of many independent factors. Attempts to esti- 
 mate this efficiency — the qualities desirable in English 
 composition — have resulted in the production of three 
 separate methods of measuring. 
 
 The first method is that of the Hillegas Scale of mixed 
 types of composition. This scale consists of a number of 
 samples of English composition representing various types 
 and ranging from very good to very poor in quality, each 
 grade in the scale being represented by but one composi- 
 tion. For example, the sample composition representing 
 one grade may be of the narration type, while that repre- 
 senting another grade may be of the description type. 
 Since the composition to be measured is compared directly 
 with the compositions in the scale, as in the Thorndike 
 Handwriting Scale, the accurate comparison of one style 
 of composition with an entirely different style, as is often 
 necessary, is exceedingly difficult. 
 
 It was to do away with this objection that the second 
 method of measurement, namely, the Harvard-Newton 
 series of four scales, was formed. These scales measure 
 efficiency in description, narration, exposition and argu- 
 mentation, respectively. 
 
 131
 
 132 Scientific Measurement 
 
 Thirdly, there is the method originated by Rice and 
 used with apparent success by Bliss and Courtis. Here 
 no attempt is made to construct an actual scale; but 
 progress in composition writing in an individual, class, or 
 school is determined by simply noting the improvement 
 shown by the individual, class, or school, in successive 
 reproductions of similar selections at intervals through- 
 out the school year. No attempt is made to express the 
 value of the composition in per cents or otherwise. It is 
 simply read, and placed in the class " Excellent," " Good," 
 " Poor," etc., on the basis of the general impression pro- 
 duced by reading it. These initial attempts are so lacking 
 in the precision for which the whole movement for stand- 
 ardization of school products stands, that they need no 
 further description. 
 
 I. HILLEGAS COMPOSITION SCALE 
 
 The "Hillegas Scale for the Measurement of Quality 
 in English Composition by Young People" consists of 
 ten sample compositions which have been arranged in 
 order of increasing merit, merit meaning that quality 
 which competent persons consider as such. These samples 
 have been assigned the following values: 0, 18, 26, 37, 
 47, 58, 67, 77, 83, and 93, respectively. These values are 
 not based on the ordinary percentage system used in 
 grading and should not be confused with such per cents. 
 Instead, each one of the values represents the number of 
 units of quality possessed by the composition to which it 
 is attached. Thus, the composition rated 93 is approxi- 
 mately twice as good as the one rated 47, while the one 
 rated 18 is approximately half as good as the one rated 37.
 
 Composition Scales 133 
 
 Dear Sir : I write to say that it aint a square deal Schools 
 is I say they is I went to a school, red and gree green and 
 brown aint it hito bit I say he don't know his business not 
 today nor yesterday and you know it and I want Jennie to 
 get me out. 
 
 18 
 
 the book I refer to reach is Ichabod Crane, it is an grate 
 book and I like to rede it. Ichabod Crame was a man and 
 a man wrote a book and it is called Ichabod Crane i like it 
 because the man called it ichabod crane when I read it for it 
 is such a great book. 
 
 26 
 
 Advantage evils are things of tyranny and there are many 
 advantage evils. One thing is that when they opress the 
 people they suffer awful I think it is a terriable thing when 
 they say that you can be hanged down or trodden down 
 without mercy and the tyranny does what they want there 
 was tyrans in the revolutionary war and so the throwed off 
 the yok. 
 
 37 
 
 Sulla as a Tyrant 
 
 When Sulla came back from his conquest Marius had put 
 himself consul so sulla with the army he had with him in his 
 conquest seized the government for Marius and put himself 
 in consul and had a list of his enemys printy and the men 
 whoes names were on this list we beheaded.
 
 134 Scientific Measurement 
 
 47 
 
 De Quincy 
 
 First : De Quincys mother was a beautiful woman and 
 through her De Quincy inhereted much of his genius. 
 His running away from school enfluenced him much as he 
 roamed through the woods, valleys and his mind became very 
 meditative. 
 
 The greatest ennuence of De Quincy's life was the opium 
 habit. If it was not for this habit it is doubtful whether 
 we would now be reading his writings. 
 
 His companions during his college course and even before 
 that time were great enfluences. The surroundings of De 
 Quincy were enfluences. Not only De Quincy's habit of 
 opium but other habits which were peculiar to his life. 
 
 His marriage to the woman which he did not especially 
 care for. 
 
 The many well educated and noteworthy friends of De 
 
 Quincy. 
 
 58 
 
 Fluellen 
 
 The passages given show the following characteristic of 
 Fluellen : his inclination to brag, his professed knowledge of 
 History, his complaining character, his great patriotism, 
 pride of his leader, admired honesty, revengeful, love of fun 
 and punishment of those who deserve it. 
 
 67 
 
 Ichabod Crane 
 
 Ichabod Crane was a schoolmaster in a place called Sleepy 
 Hollow. He was tall and slim with broad shoulders, long 
 arms that dangled far below his coat sleeves. His feet 
 looked as if they might easily have been used for shovels. 
 His nose was long and his entire frame was most looely hung 
 to-gether.
 
 Composition Scales 135 
 
 77 
 
 Going Down with Victory 
 
 As we road down Lombard Street, we saw flags waving 
 from nearly every window. I surely felt proud that day to 
 be the driver of the gaily decorated coach. Again and again 
 we were cheered as we drove slowly to the postmasters, to 
 await the coming of his majestie's mail. There wasn't one 
 of the gaily bedecked coaches that could have compared 
 with ours, in my estimation. So with waving flags and 
 fluttering hearts we waited for the coming of the mail and the 
 expected tidings of victory. 
 
 When at last it did arrive the postmaster began to quickly 
 sort the bundles, we waited anxiously. Immediately upon 
 receiving our bundles, I lashed the horses and they responded 
 with a jump. Out into the country we drove at reckless 
 speed — everywhere spreading like wildfire the news, "Vic- 
 tory!" The exileration that we all felt was shared with 
 the horses. Up and down grade and over bridges, we drove 
 at breakneck speed and spreading the news at every hamlet 
 with that one cry "Victory !" When at last we were back 
 home again, it was with the hope that we should have an- 
 other ride some day with "Victory." 
 
 83 
 
 Venus of Melos 
 
 In looking at this statute we think, not of wisdom, or 
 power, or force, but just of beauty. She stands resting the 
 weight of her body on one foot, and advancing the other (left) 
 with knee bent. The posture causes the figure to swav 
 slightly to one side, describing a fine curved line. The 
 lower limbs are draped but the upper part of the body is un- 
 covered. (The unfortunate loss of the statute's arms pre-
 
 136 Scientific Measurement 
 
 vents a positive knowledge of its original attitude). The 
 eyes are partly closed, having something of a dreamy lan- 
 gour. The nose is perfectly cut, the mouth and chin are 
 moulded in adorable curves. Yet to say that every feature 
 is of faultless perfection is but cold praise. No analysis can 
 convey the sense of her peerless beauty. 
 
 93 
 
 A Foreigner's Tribute to Joan of Arc 
 
 Joan of Arc, worn out by the suffering that was thrust 
 upon her, nethertheless appeared with a brave mien before 
 the Bishop of Beauvais. She knew, had always known that 
 she must die when her mission was fulfilled and death held 
 no terrors for her. To all the bishop's questions she answered 
 firmly and without hesitation. The bishop failed to confuse 
 her for heresy, bidding her recant if she would live. She 
 refused and was lead to prison, from there to death. 
 
 While the flames were writhing around her she bade the 
 old bishop who stood by her to move away or he would be 
 injured. Her last thought was of others and De Quincy 
 says, that recant was no more in her mind than on her lips. 
 She died as she lived, with a prayer on her lips, and listening 
 to the voices that had whispered to her so often. 
 
 The heroism of Joan of Arc was wonderful. We do not 
 know what form her great patriotism took or how far it 
 really led her. She spoke of hearing voices and seeing visions. 
 We only know that she resolved to save her country, know- 
 ing though she did so, it would cost her her life. Yet she 
 never hesitated. She was uneducated save for the lessons 
 taught her by nature. Yet she led armies and crowned the 
 dauphin, king of France. She was only a girl, yet she could 
 silence a great bishop by words that came from her heart 
 and from her faith. She was only a woman, yet she could 
 die as bravely as any martyr who had gone before. 
 
 This scale is reproduced by the courtesy of Dr. M. B. Hillegas.
 
 Composition Scales 137 
 
 The scale was derived in the following manner. The 
 first step taken was the collection from various sources of 
 about 7000 English compositions ranging from the very- 
 poorest to the best work done in the elementary and high 
 schools. After these compositions had each been given 
 a number from 1 to 7000, they were roughly graded by 
 Hillegas and an assistant into ten classes, and from these 
 ten classes 75 samples were selected. In order to have 
 samples at both extremes of the scale, some artificial ones 
 were supplied. Those placed at the zero end of the scale 
 were conscious efforts by adults to write very poor Eng- 
 lish, while those placed at the one hundred end were 
 obtained from youthful writings of certain literary geniuses 
 and from the work of some college freshmen. As aug- 
 mented, the set consisted of 83 samples varying from the 
 poorest to the best by small degrees of quality. That 
 the character of the handwriting might not influence the 
 judges, all the samples were typewritten and mimeo- 
 graphed. 
 
 Separate sets of these samples were then sent to about 
 100 individuals, who were asked to arrange the samples 
 in the order of their merit as specimens of English com- 
 position, calling the poorest specimen No. 1, the next, 
 No. 2, and so on. Owing to the small number of judg- 
 ments it was not possible to establish the position of any 
 one sample with reasonable accuracy, but those samples 
 that were of about equal merit were indicated. This re- 
 sulted in the selection of a smaller set which still con- 
 tained all the important steps in quality from the worst 
 to the best. 
 
 This smaller set, comprising 27 samples, was selected 
 by taking successively each of the samples in the larger 
 group that about 75% of the judges had agreed was 
 better than the last one selected. This percentage of 
 judgments was taken for statistical reasons which will be 
 explained later. Where large differences in merit existed
 
 138 Scientific Measurement 
 
 between two successive samples, new samples, judged by 
 a number of individuals as ranging in merit between 
 them, were introduced. 
 
 Then, as with the first set of samples, more than 100 of 
 these sets consisting of 27 samples were mailed to com- 
 petent critics of English literature, such as teachers, 
 authors, and literary workers, with the request to rank 
 them in order of literary merit. When 75 replies had 
 been received, the results were tabulated as in the case 
 of the first set. Meantime, the judgments of 41 indi- 
 viduals especially competent to judge merit in English 
 composition writing were secured to use as a check on 
 the others. The examination of the results from this 
 second set showed the necessity of adding two more 
 samples to the set. This was done, making 29 samples 
 in all. 
 
 After one or the other of the two sets, to which 21 of 
 the samples were common, had been judged by about 
 200 individuals, it was decided to make the scale. The 
 first thing necessary was to locate a zero point. This 
 point was to be represented by a sample which possessed 
 absolutely no merit as an English composition. It was 
 chosen on the basis of the judgments of 28 qualified indi- 
 viduals. When the result of these judgments was tabu- 
 lated, it was found that just one-half of them considered 
 such a point as below sample 580 and one-half as above 
 it, and so sample 580 was taken as the zero point on the 
 scale. 
 
 The ten samples chosen for the scale were selected on 
 the principle of equally often noticed differences, which 
 is as follows : Differences that are equally often noticed 
 are equal (unless always or never noticed). Thus, if in 
 a set of samples, a, b, c, d, etc., it was found that a was 
 judged better than b, just as often as b was judged better 
 than c, and so on, samples a, b, c, d, etc., would constitute 
 a scale of equal steps. To put the case more concretely,
 
 Composition Scales 139 
 
 if in an essay contest, essay A was judged better than 
 essay B in 75% of the judgments, essay B was judged 
 better than essay C in the same number of judgments, 
 and so on, it is readily seen that the differences in quality 
 between essays A, B, C, etc., are equal because the same 
 number of individuals noticed this difference. Similarly, 
 as a result of all the comparisons made of the sample 
 compositions, the result was approximately as follows : 
 
 Sample 18 was judged better than sample in 75% of 
 the judgments. 
 
 Sample 26 was judged better than sample 18 in 75% of 
 the judgments. 
 
 Sample 37 was judged better than sample 26 in 75% of 
 the judgments, and so on for samples 47, 58 and 94. 
 
 Thus in samples 18, 26, 37, etc., we have the successive 
 steps of a scale, steps that are equal inasmuch as they 
 represent differences that are equally often noticed. 
 
 Why the opinion of 75% of the judges was taken as 
 the unit of value, instead of some other per cent, may 
 probably be better understood if the following case is 
 considered. If, in comparing the ability of two states- 
 men, say Gladstone and Bismarck, 50% of the judges 
 claim Gladstone to have possessed the greater ability, 
 while 50% claim the same for Bismarck, it may safely be 
 assumed that they possessed about equal ability. If, 
 however, 60% of the judges believe Gladstone to have 
 been the more efficient, the chances are that Gladstone 
 was probably slightly more capable than Bismarck. As 
 the percentage of judgments favoring Gladstone increases, 
 the chances are shown to be greater that Gladstone had 
 the superior ability, and when 100% of the judges believe 
 him to have surpassed Bismarck it may safely be assumed 
 that such was actually the case. Similarly, in the present 
 case if 75% of the judges say that a given sample is 
 better than another given sample, we may be reasonably 
 sure that such is the case.
 
 140 Scientific Measurement 
 
 The value of any English composition may be obtained 
 by placing it alongside the samples in the scale and decid- 
 ing which it is most nearly like in quality. By having 
 other judges measure it, each being in ignorance of the 
 judgment of the others, or, if this is not practicable, by 
 rating the sample two or three times, a very accurate 
 measure of it may be secured. For example, if the com- 
 position seems to be very similar in quality to sample 77, 
 then it is marked 77. If it seems to lie between samples 
 77 and 83, it should be given a value between 77 and 83, 
 as 79 or 81, according to which sample the specimen more 
 nearly resembles. 
 
 II. HARVARD-NEWTON SCALES 
 
 An experiment with the Hillegas Scale in the public 
 schools of Newton, Massachusetts, led the school authori- 
 ties of that city to believe that it possessed several in- 
 herent defects. They maintained that since the scale 
 provides one, and only one, type of composition for each 
 one of the grades, the type of one grade differing entirely 
 from that of the next (that is, grade A in the scale is 
 represented by one type of composition, grade B, by an- 
 other, and so on), it was difficult or impossible to com- 
 pare the work of one type of composition, narration, for 
 example, with that of another type, like description. 
 Moreover, they claimed the sample compositions were 
 not typical of efficient school work. An attempt to 
 remedy these defects resulted in the Harvard-Newton 
 series of scales, the general nature of which will be de- 
 scribed before the construction of the scale is discussed 
 in detail. This objective measure is the outcome prin- 
 cipally of the cooperation of Ballou, and the teachers of 
 the Boston and Newton public school systems. 
 
 It consists of four separate scales to measure the four 
 different forms of composition in the eighth grade ; namely, 
 description, narration, argumentation, and exposition.
 
 Composition Scales 141 
 
 Each scale in the series is composed of six compositions, 
 actually written by eighth grade pupils ; thus each scale 
 possesses the same qualities that it is designed to measure. 
 These sample compositions range by approximately equal 
 steps from the best to the poorest work which is likely 
 to be done in the eighth grade, and each of them has 
 been assigned a letter and a percentage valuation in con- 
 formity with the current practice in grading. "A" rep- 
 resents the conventional value of 95%; "B" that of 
 85%; "C" of 75%; and so on. In this way sample "A" 
 is fairly representative of all compositions whose value 
 would seem to lie somewhere between 90% and 100%; 
 sample "B", of all those whose value would seem to lie 
 between 80% and 90%, and so on. Each sample com- 
 position is accompanied with a short description of its 
 merits and defects, and it is compared with the next 
 higher and lower compositions in the scale. These de- 
 scriptions and comparisons were written by the teachers 
 who helped to make the scale and expected to use it. 
 Without some such guiding material, it is doubtful 
 whether those who use the scale would see the same 
 merits and defects in a composition as those who made 
 the scale, and, unless this was the case, little advantage 
 would be derived from its use. The general nature of 
 the four scales may readily be seen from the one — the 
 description scale — which follows.
 
 142 Scientific Measurement 
 
 THE COMPLETED DESCRIPTION SCALE 
 No. 1. "A" GRADE COMPOSITION. VALUE, 94.6% 
 
 A Storm in a Fishing Village 
 
 It was a cold damp day in November. The sky was a 
 heavy leaden color. In the east a black line stretched 
 across it foretelling the coming of a storm. The houses 
 across the way were dismal shadows, — flat, cold, heart- 
 
 5 less. A piercing chill penetrated to the bone. The rattle 
 of a grocer's cart or the clatter of a horse's hoofs, seemed 
 cold. The pedestrians were all clothed in black, or else 
 the feeble light made them seem so, and they were cold 
 — everything was cold, cold, cold. An awful lonliness 
 
 10 pervaded all. 
 
 The black line in the east had grown into a cloud and 
 was coming nearer, nearer, over the sea. Suddenly a gust 
 of wind shook the very foundations of the houses, — an- 
 other, and then a continuous blowing. The howling was 
 
 15 horrible. Great sheets of foam were blown into the 
 streets, — here and there a piece of wreckage hurled itself 
 against a cottage. Fishermen's wives hurried down the 
 narrow streets to the shore, straining their eyes for any 
 sign of a wreck. Old seamen looked at the roaring sea 
 
 20 and shook their heads. 
 
 By this time the black cloud had engulfed the sky. The 
 day was like night, although it was not yet noon. Boys 
 ran about with torches which were immediately extin- 
 guished, and the roaring called to mind the last day at 
 
 25 Pompeii. 
 
 Rain had begun to descend. At first only drops fell 
 on the hardened faces of old mariners, and on the pale 
 countenances of wives, mingling with the drops already 
 there. But soon great sheets fell, forcing the people in- 
 
 30 doors, to the poor shelter afforded by the groaning houses.
 
 Composition Scales 143 
 
 For about an hour the storm continued thus, then by- 
 degrees the wind lessened, though the rain still fell, and 
 the ocean thundered. But soon the rain also slowly- 
 stopped and the roaring ceased. The black cloud rolled 
 slowly away, leaving the tardy sun to shine on the drenched 35 
 town and the great piles of wreckage on the shore. 
 
 Merits 
 
 This theme ranks high because the writer has a clear picture of the 
 scene and has used words and phrases that bring the details of this 
 picture clearly before the reader. There are good color images in 
 such expressions as leaden, a black line, great sheets of foam, the day 
 was like night, and the sun shining on the drenched town. Sound effects 
 are strikingly brought out by such phrases as the rattle of a grocer's 
 cart, the howling, the wreckage hurled against the cottage, the roaring sea, 
 and the thundering ocean. The sensation of dreariness and chill is 
 conveyed by the repetition of the word cold. The confusion caused 
 by the storm is reflected in the anxious look of the wives of the fisher- 
 men. A further human touch is added in the mention of such details 
 as the extinguished torches carried by the boys and the drops of rain fall- 
 ing upon the hardened faces of the old mariners. All these enumera- 
 tions fittingly combine to produce a tone of coldness, desolation, and 
 anxiety. The details are told in their natural sequences. This 
 chronological arrangement has helped the writer to keep safely to his 
 main point and effectively connect the details with each other. 
 
 Defects 
 
 The repetition of the word cold, while effective in bringing out the 
 sensation, is somewhat artificial. Loneliness (line 9), is misspelled; 
 a semicolon should supplant the comma in line 8. Omit the comma 
 in line 6. 
 
 Comparison 
 
 The theme is superior to No. 2 in its richness of imagery, its wealth 
 of details, its depth of feeling, its maturity of style (seen in the sen- 
 tence-structure and the vocabulary), and in its mastery of mechanical 
 forms.
 
 144 Scientific Measurement 
 
 No. 2. "B" GRADE COMPOSITION. VALUE, 83.5% 
 
 Grandmother 
 
 In front of the open fireplace in a large armchair there 
 sits our old Granny. She is old and feeble. Her hair is 
 snow-white and over her head a little white cap is care- 
 fully tied. Her face is full of wrinkles and her keen blue 
 5 eyes sparkle through a pair of glasses which she has on 
 her nose. 
 
 She has a shawl thrown over her shoulders and she also 
 wears a thick black skirt. On her feet can be seen a pair 
 of soft slippers which she prizes very much because they 
 10 were given her for a Christmas present. 
 
 As you know Grannies always like to be busy our 
 Granny is busy knitting gloves. Her hands go to and fro. 
 She will keep on working until her knitting is done. Now 
 that it is done she carefully folds her work and packs it 
 15 into her workbasket. Then she trots upstairs to bed and 
 oh, how lonesome it is when our dear Granny is gone 
 
 from the room. 
 
 Merits 
 
 The merits of this composition are : (1) the clear and pleasing im- 
 pression obtained; (2) the happy choice of details and the logical 
 sequence of their arrangement ; (3) the sympathetic treatment of the 
 subject — for example, bits of sentiment seen in the grandmother's 
 attachment to the slippers, and the loneliness felt when she goes to 
 her room; (4) the interesting introductory sentence; and (5) the 
 mechanical accuracy. 
 
 Defects 
 
 The defects are: (1) the rather monotonous sentence structure, 
 and (2) the childish vocabulary. 
 
 Comparison 
 
 To justify its place in the scale, note : (1) that in No. 1 there is 
 successfully treated a much more difficult subject; (2) there is a 
 greater power of imagination; and (3) there is a greater variety of 
 sentence structure and a richer vocabulary.
 
 Composition Scales 145 
 
 No. 3. "C" GRADE COMPOSITION. VALUE, 76.1% 
 
 A Mansion 
 
 As you look across the road you will first see a long 
 private avenue or walk. 
 
 It is in the summer, and on each side of this long walk 
 are some beautiful, stately elms. They are hundreds of 
 years old and they have done their duty for as many, 5 
 years, shading the walk from the noon sun. 
 
 Cross the road and you will see if you look up the 
 avenue, a beautiful mansion. It is a colonial house and 
 four large pillars are upholding the roof. A piazza runs 
 along three sides of the house. 10 
 
 Near the house is a tennis court where for years the 
 occupants of the mansion have passed many an hour. 
 
 Let us enter the mansion. It is a beautiful cool place, 
 although dark. As we enter we see large psalms on each 
 side of the entrance. On the floors are old oriental rugs 15 
 which have been handed down for generations. In the 
 parlor is a harp, and on the walls are the portraits of the 
 ancestors. In all, it is a beautiful place. 
 
 Merits 
 
 The writer of this theme has presented a clear though conven- 
 tional picture. Although he changes his point of view several times, 
 he has attempted to put his readers into the best positions to see the 
 mansion. The choice of words is fair. Such details as the stately 
 elms, the oriental rugs, the harp, and the portraits are well selected. 
 Only one mistake in spelling occurs (line 14). 
 
 Defects 
 
 There are, however, too many paragraphs for such a short theme. 
 Constant repetition of the pronoun you, and of the words beautiful 
 and mansion give an impression of monotony and of limited vocabu- 
 lary. The pupil has evidently a definite place in mind, but has not 
 suggested the spirit of the scene, as has the writer of No. 2.
 
 146 Scientific Measurement 
 
 Comparison 
 
 The composition deserves its place in the scale above No. 4 be- 
 cause of better sentence structure and more orderly arrangement. 
 It is inferior to No. 2 on account of its somewhat prosaic tone and its 
 constantly changing point of view. 
 
 No. 4. "D" GRADE COMPOSITION. VALUE, 66.6% 
 
 The Lake at Sunrise 
 
 In the Mountains of Pennsylvania there is a lake. 
 On one side of the lake is a boat landing, at which a 
 dozen or more boats are tied up. On this boat landing 
 one may stand and look up the lake, at sunrise, and see 
 5 the sun peering up over the top of the mountains and 
 shinning on the water. Then a King Fisher flies down the 
 lake making his cheerful noise, instantly, all the other 
 birds begin to chirp as if their life depended on it. 
 
 Looking across the lake one would see numerous wells 
 
 10 and coves backed up by woods from which comes the chirp 
 
 of the birds. Hearing the explosions of cylinders we look 
 
 to see where in comes from and find a pumphouse that 
 
 keeps the lake supplied with water. 
 
 Looking down the lake over the dam to the ice house 
 15 with the roof sparkling with. On the roof of the house a 
 hawk is sitting adding his clear whistle to noise of other 
 birds. 
 
 Looking around to the woods, at our back, with an old 
 oil well in front of them. The birds flying from the woods 
 20 in flocks, and far away from the hills comes the sound of 
 the of Italians singing. 
 
 Merits 
 
 The writer has seen and heard concrete details and has re-created 
 his images clearly. He has tried, too, to make his point of view 
 obvious to the reader. His vocabulary is adequate.
 
 Composition Scales 147 
 
 Defects 
 
 As a description the composition fails because there is no unified 
 picture of the lake. The selected details, clear in themselves, tend 
 to distract rather than center the interest. There are numerous 
 mechanical errors : there should be no commas after lake or sunrise 
 (line 4) ; shining (line 6) is misspelled ; there should be a period after 
 noise (line 7), and no comma after instantly (line 7), which should 
 commence with a capital ; in (line 12) is not correct ; the groups of 
 words in lines 14, 15, and lines 17, 18 do not make sentences; 
 the word the is omitted before noise (line 16) and the word are before 
 flying (line 18). 
 
 Comparison 
 
 The theme merits its rank in the scale by superiority in spelling, 
 paragraphing, and maturity of thought. It does not, on the other 
 hand, show equal mastery in the fine details, the discriminating vocabu- 
 lary, and in the ability to stick to the point. The sentence-sense is 
 faulty. 
 
 No. 5. "E" GRADE COMPOSITION. VALUE, 55.4% 
 
 A Light House 
 
 A description of a light house is quite interesting. 
 
 First a light house is generally situated on a mass of 
 rocks in the ocean or on some great lake. And then to 
 get into a light house is a question. Some times you have 
 to climb to the top on a steal ladder, and again you only 5 
 have to go half way up and you find sort of a steal porch, 
 which is very strong with a door in the side of the light 
 house. On the very top of the light there is generally 
 two or three life boats in case of accidents. In side there 
 is an enormous light which flashes every two minutes and 10 
 sometimes more often it depends holy on the weather. 
 The man himself has very favorable sleeping quarter and 
 food it is a very lonely life except when you have a man 
 with you. Sometimes they play cards all day long until 
 it is time to fix the lights and then they are very busy. 15
 
 148 Scientific Measurement 
 
 Merits 
 
 The merits of this theme are: (1) the evident spirit of faithful 
 accuracy ; and (2) a successful use of certain simple words, — such as 
 mass of rocks, enormous light, and lonely life. 
 
 Defects 
 
 Many obvious defects warrant its low position in the scale. The 
 pupil was asked to write a description. After announcing his pur- 
 pose to do this, he writes an exposition, or explanation of lighthouses 
 in general. The first sentence of the theme is worthless, contributing 
 nothing toward the development of the subject. It should be omitted. 
 The paragraph is full of misspelled words and grammatical slips; 
 steal, in side, holy, some times, sleeping quarter. The most striking 
 weakness of the work is the loose and rambling form of the sentences, 
 indicating indefinite thought. "Run-on" sentences are found in 
 lines 9-13. No attempt has been made to establish a point of view. 
 On this account, and because of a lack of vivid words, the passage is 
 dead and colorless. 
 
 Comparison 
 
 The composition is placed above No. 6 because it contains fewer 
 mechanical errors. 
 
 NO. 6. "F" GRADE COMPOSITION. VALUE, 44.9% 
 
 A Scene on the Prairies 
 
 Along a large plain in the west with mountains on all 
 sides. The sun was just sinking behind the mountains. 
 Some trappers were on the plain just about to get their 
 supper. They had one tend because there was just three 
 
 5 of them. Beside their tent tripled a little spring. After 
 the three trappers had eating there supper they sat down 
 by the fire because it had growing dark. All of a sudden 
 a bunch of Indain's came riding up. When they came 
 near they fired of their guns and disappered in the dark- 
 
 10 ness and the trappers turned into camp leaving one a the 
 trappers on gaurd. 
 
 a
 
 Composition Scales 149 
 
 Merits 
 
 The commendable features of this composition are directness, 
 simplicity, and a logical arrangement of details. The writer passes 
 from the general to the specific in a natural manner. In spite of a 
 change in the point of view in the last two sentences, the paragraph, 
 as a whole, makes a clear picture. 
 
 Defects 
 
 Blunders in grammar and in spelling, lack of sentence-sense, and 
 short, childish sentences make the rating of the composition necessarily 
 very low. Such errors as tend for tent, tripled for trickled, eating for 
 eaten, growing for grown, and the misspelling of Indians indicate 
 either hasty, careless work, or slovenly habits of enunciation. 
 
 Comparison 
 
 Compared with the descriptions of the storm and of grandmother, 
 the short sentences here show immaturity and weakness rather than 
 skill or force. With a large amount of correcting of mechanical de- 
 tails, but with very little revising as a whole, this composition would 
 be superior to No. 5. 
 
 The scales and tables in this section are reproduced by the courtesy 
 of Dr. F. W. Ballou.
 
 150 Scientific Measurement 
 
 EFFECT OF USING THE SCALE 
 
 An initial experiment in the use of the description scale 
 was made in Arlington and Boston. Eighth grade teachers 
 and elementary school principals in these two cities graded 
 a set of twenty-five eighth grade compositions secured for 
 this purpose, both without the use of the scale and with 
 it. With the use of the scale the results showed a reduc- 
 tion in the extreme variation of judgments; that is, no 
 two teachers were quite so widely divergent as before. 
 The average variation was also less. But in this matter 
 neither the average nor the extreme variation is the most 
 important consideration. Far more important is the 
 effect which the use of the scale has on the grading of each 
 individual teacher. To ascertain this is obviously a com- 
 plicated matter, and it requires more time than has been 
 thus far at our disposal. This phase of the problem will 
 be the subject of further investigation. 
 
 The compositions used in the scale were selected from 
 a large number written by the eighth grade pupils of 
 Newton as a part of their regular school work. Each 
 pupil was given his choice among several topics of descrip- 
 tion, narration, exposition, and argumentation, suggested 
 by himself or the teacher, and was required to write a 
 composition of about a page in length. Time for prepara- 
 tion and correction was allowed. Thus, these composi- 
 tions represented the best unaided writing of the indi- 
 vidual children in the eighth grade of that particular city. 
 Then a selection from all these compositions was made 
 by the individual eighth grade teachers. This selection 
 included at least 25% of all the compositions written in a 
 particular class and was made with the view of securing 
 compositions representing all degrees of ability in that 
 class. The compositions were then numerically graded 
 by the eighth grade teacher and the principal, inde- 
 pendently. To be sure of securing compositions deserv-
 
 Composition Scales 151 
 
 ing the highest grade of merit, namely, "A" or 95%, 
 each school, in addition, sent in from one to three of its 
 "best" compositions in all four types of writing, as 
 judged by the teacher and principal. Twenty-five samples 
 of each one of the four types of composition — description, 
 narration, exposition, and argumentation — seemed a 
 sufficient number from which to select the six composi- 
 tions to be used in the final construction of each one 
 of the four scales. Twenty-five samples, then, of 
 each type were selected on the basis of the preliminary 
 grading given the compositions by the teachers and prin- 
 cipals and on the judgment of Ballou, director of the 
 experiment. 
 
 To eliminate any possible influence of handwriting 
 these samples were typewritten and mimeographed. 
 Then one set, consisting of 25 samples of each of the four 
 types of composition, was sent to each of the eighth 
 grade teachers and principals, 25 in all, with instructions 
 
 (1) to grade each of the compositions independently and 
 
 (2) to rank each in the order of its merit. 
 
 Because of the probability that 95% rather than 100% 
 would represent the highest degree of efficiency in com- 
 position writing in the eighth grade, and because it was 
 desirable that each reader should start from the same 
 point in marking the compositions, the teachers were 
 asked to give 95% to the best compositions. Although 
 no lower limit was fixed, 40% was intended to be that 
 limit ; for compositions worth less than that were not to 
 be furnished by the schools for the experiment. 
 
 As already stated each composition was graded by 25 
 teachers, and, when the marks came in, five things were 
 noted with regard to each of them : 
 
 (1) Its average mark (found by dividing the sum of all 
 the marks by 25). 
 
 (2) Its median mark (found by ranging all the marks 
 given it in order from the highest to the lowest and taking
 
 152 
 
 Scientific Measurement 
 
 the middle one). (This is easier to find than the average 
 and for many purposes it is better.) 
 
 (3) The highest mark given it. 
 
 (4) The lowest mark given it. 
 
 (5) The difference between these two, which is the 
 maximum variation in the marking of these particular 
 compositions. 
 
 Marks Given to the Twenty-five Compositions 
 
 Composi- 
 
 Highest 
 
 Lowest 
 
 Maximum 
 
 Mean or Aver- 
 
 Median 1 
 
 tion 
 Number 
 
 Grade 
 
 Grade 
 
 Variation 
 
 age Grade 
 
 Grade 
 
 1 
 
 95 
 
 68 
 
 27 
 
 91.9 
 
 83.0 
 
 2 
 
 90 
 
 64 
 
 26 
 
 80.0 
 
 80.0 
 
 3 
 
 50 
 
 30 
 
 20 
 
 42.7 
 
 41.0 
 
 4 
 
 94 
 
 63 
 
 31 
 
 84.3 
 
 85.5 
 
 5 
 
 78 
 
 50 
 
 28 
 
 61.1 
 
 60.0 
 
 6 
 
 88 
 
 50 
 
 38 
 
 69.4 
 
 69.5 
 
 7 
 
 80 
 
 40 
 
 40 
 
 63.5 
 
 65.0 
 
 8 
 
 95 
 
 52 
 
 43 
 
 82.3 
 
 85.0 
 
 9 
 
 75 
 
 40 
 
 35 
 
 56.1 
 
 58.5 
 
 10 
 
 95 
 
 90 
 
 5 
 
 94.5 
 
 95.0 
 
 11 
 
 65 
 
 40 
 
 25 
 
 49.5 
 
 49.5 
 
 12 
 
 75 
 
 42 
 
 33 
 
 59.9 
 
 60.0 
 
 13 
 
 95 
 
 71 
 
 24 
 
 83.7 
 
 85.0 
 
 14 
 
 76 
 
 40 
 
 36 
 
 55.4 
 
 53.5 
 
 15 
 
 95 
 
 80 
 
 15 
 
 89.6 
 
 90.0 
 
 16 
 
 92 
 
 68 
 
 24 
 
 78.2 
 
 78.5 
 
 17 
 
 93 
 
 63 
 
 30 
 
 81.0 
 
 81.5 
 
 18 
 
 90 
 
 60 
 
 30 
 
 79.9 
 
 75.0 
 
 19 
 
 92 
 
 60 
 
 32 
 
 79.6 
 
 80.0 
 
 20 
 
 92 
 
 70 
 
 22 
 
 82.7 
 
 85.0 
 
 21 
 
 89 
 
 54 
 
 35 
 
 76.1 
 
 77.0 
 
 22 
 
 86 
 
 47 
 
 39 
 
 66.6 
 
 66.5 
 
 23 
 
 74 
 
 40 
 
 34 
 
 55.4 
 
 57.5 
 
 24 
 
 73 
 
 30 
 
 43 
 
 48.9 
 
 48.0 
 
 25 
 
 62 
 
 20 
 
 42 
 
 44.9 
 
 45.0 
 
 As a check on the results of the gradings, the returns 
 from the rankings were also tabulated and the same 
 items noted as in the case of the grades. 
 
 1 "Median grade" is the grade in the series of grades above which 
 and below which there is an equal number of grades.
 
 Composition Scales 153 
 
 After the various items in both grading and ranking 
 had been recorded for each composition, using these data 
 as a basis, it was necessary to choose the compositions 
 best fitted to have a place in the scale. It is obvious 
 that compositions about which there was most agreement 
 in judgment on the part of the teachers, both as to rank 
 and grade — that is, compositions with low maximum 
 variations — were most desirable ; furthermore, since it 
 was the intention of the authors of the scale that the six 
 compositions selected should represent 95%, 85%, 75%, 
 65%, 55% and 45%, respectively, in choosing the composi- 
 tions for the scale they accordingly selected those whose 
 average and median marks came nearest those require- 
 ments. 
 
 In short, in constructing the scale there were no fixed 
 requirements set. The compositions selected were those 
 about which there was the least disagreement as to merit 
 and whose marks approximated those desired in the scale. 
 After the six compositions had been selected on this 
 basis, the teachers were asked to point out in a brief para- 
 graph the merits and defects of each of the compositions. 
 These paragraphs were carefully studied and compared 
 by a committee who, acting under expert advice, put the 
 various criticisms into the form shown in the scale already 
 presented. 
 
 The method of using this scale is very simple. The 
 composition to be measured is compared directly with 
 those in the appropriate scale — description, narration, 
 etc. — and its value determined in terms of the marks 
 assigned to the sample composition which it most nearly 
 approaches in quality. Thus a descriptive composition 
 is placed alongside the compositions in the description 
 scale, a narrative composition alongside the compositions 
 in the narration scale, etc. If the composition to be 
 measured seems to possess the same qualities as a given 
 composition in the scale — say the composition represent-
 
 154 Scientific Measurement 
 
 ing grade "B" in the description scale — then it is 
 assigned the same value as that composition, namely 
 grade "B" or 83.5%. If its value seems to lie some- 
 where between two grades on the scale as represented by- 
 two compositions, say "A" (94.6%) and "B" (83.5%), 
 the examiner can determine its value as precisely as he 
 pleases according to its apparent distance below the one 
 and above the other. 
 
 In spite of the difficulty of comparing a sample of com- 
 position writing of one type with a sample of another 
 type, as is necessary in using the Hillegas Scale, in actual 
 practice the Hillegas Scale has on the whole been used to 
 greater advantage than the Harvard-Newton Scale. 
 This has been due chiefly to the fact that the field in 
 which the former may be used — the elementary grades 
 and high school — is not as limited as that of the latter, 
 which is confined to the eighth grade. However, for 
 eighth grade measurements the Harvard-Newton Scale 
 may obviously be used to better advantage. 
 
 The teacher may obtain the Hillegas Scale by sending 
 to Teachers College, Columbia University, New York. 
 To recapitulate, all that need be done in using it is to 
 slide the composition to be measured along the scale — 
 as in the case of the handwriting scales — beginning with 
 the sample marked 0, until a sample is reached on the 
 scale to which the specimen to be measured most closely 
 corresponds in quality. As has been said, the former 
 may be of an entirely different type from the latter. 
 The composition to be measured is then given the same 
 value as the one on the scale to which it is most similar 
 in quality. That is, if it appears to be very like the com- 
 position marked 77 it is given the value 77. If it seems to 
 be better than composition 77 but not so good as the 
 next composition in the scale, number 83, it is given a 
 value somewhere between 77 and 83 such as 79 or 81. 
 
 Teachers of the eighth grade may obtain the Harvard-
 
 Composition Scales 155 
 
 Newton Scale by sending to The Harvard University- 
 Press, Boston, Mass. In using it, a descriptive composi- 
 tion is measured by comparing it with the sample com- 
 positions on the description scale, a narrative composi- 
 tion, by comparing it with samples in the narration scale, 
 etc. 
 
 Whichever scale is used, in obtaining the compositions 
 to be measured, the teacher must see first of all that the 
 same amount of time for writing is allowed to all the 
 pupils and, secondly, that the same subject is given to 
 all to write upon. Even in thus making the conditions 
 under which the compositions are obtained as objectively 
 uniform as possible, it is apparent that certain subjective 
 influences, such as interest for example, which cannot be 
 eliminated, are bound to affect the result. Furthermore, 
 within the same class there will be the widest difference 
 in the amount of material written. 
 
 While it is evident that in disregarding these two fac- 
 tors the scales are not complete as adequate measures of 
 composition writing, still they are of great value ; for by 
 their use the composition work of any grade, school, or 
 system of schools in any part of the country may be 
 compared with that of any other, and the results of dif- 
 ferent methods of instruction or of other conditions ascer- 
 tained and utilized. Moreover, there is so intimate a 
 relation between the successful use of oral and written 
 language and intelligence that an objective standard which 
 accurately measures ability in the use of language also 
 measures, to a certain extent, the possession of mental 
 ability in general. In the writing of English composi- 
 tion, whatever its type, children are compelled, or should 
 be compelled, above everything else to make themselves 
 clear, and, by the use of a uniform standard of judgment, 
 the growth of reason itself, from grade to grade, may be 
 followed and subnormal or supernormal children detected. 
 Then, too, the difference shown by the same child in the
 
 156 Scientific Measurement 
 
 various types of composition may give a fair idea of his 
 individuality. Increased knowledge of the various types 
 of pupils with which the school has to deal will naturally 
 lead to greater variety in teaching and correspondingly 
 better results with the children. Any such educational 
 progress, however, will come not as an expression of mere 
 opinion, but as the result of scientifically determined 
 educational facts obtained by the use of objective stand- 
 ards. The more scientific, yet comprehensible, are our 
 methods of investigation, the more valuable will be their 
 results. 
 
 EXERCISES 
 
 1. In what way may these scales be utilized to secure a very accu- 
 rate judgment of the merit of a given composition? 
 
 2. What relation seems to exist between ability in composition 
 writing and ability in other subjects in the curriculum? Between 
 ability in composition writing and general intelligence ? 
 
 3. Procure twenty compositions from various grades and get five 
 teachers to mark them on a percentage basis. What do the results 
 show regarding the reliability of such measures? 
 
 4. How would the ratings given by five teachers to twenty com- 
 positions of varying merit test the reliability of the Hillegas Scale? 
 
 5. Suppose the Composition Scale revealed a great difference in 
 the same child in the various types of composition writing, of what 
 value would this be to the teacher? 
 
 6. Obtain forty specimens of English composition from the various 
 grades. Grade these on the Hillegas Scale. Allow one month to 
 elapse and grade again. What do the results show? 
 
 7. In what type of composition writing do you think a child should 
 be most proficient? 
 
 8. Suppose a teacher discovered by the use of the scales that the 
 pupils on the whole showed far greater efficiency in one type of com- 
 position writing than in another, what should be the conclusion? 
 
 9. How would you modify the standard for composition writing 
 for your particular grade? Why? 
 
 10. What modifications would you make in it if your pupils came 
 from a foreign neighborhood?
 
 CHAPTER VII 
 COMPLETION TEST LANGUAGE SCALES — TRABUE 
 
 Suppose we consider an incomplete sentence such as 
 the following: "The . . . rises . . . the morning and 
 ... at night," where three words are omitted, the 
 place of each word being filled by a dotted line; it is a 
 simple matter for any one who is acquainted with the 
 English language to insert a word in each of these three 
 blank spaces, which will cause the sentence to make sense. 
 In the above example, these words are "sun," "in," and 
 "sets," making the sentence read : "The sun rises in the 
 morning and sets at night." The completion of sentences 
 of this kind, while not actually testing ability in English 
 composition, demands an ability very closely related to 
 what is usually called "language ability"; at any rate, 
 it involves a power to read and think about printed words 
 which has great educational significance. , 
 
 From the nature of this test it is obvious that we may 
 have sentences for completion of all degrees of difficulty. 
 While a sentence such as, "The sky . . . blue," requires 
 next to no ability in English language, a sentence such 
 as the following; "To . . . friends is always . . . the 
 .... it takes," is of sufficient difficulty to test the 
 ability of a college student. If, therefore, we could select 
 a series of incomplete sentences increasing in difficulty 
 from the first to the last, with this as a scale, we should 
 be in a position to measure the language ability of any 
 individual or group. This could be accomplished by 
 allowing a certain specified time in which to complete as 
 
 157
 
 158 Scientific Measurement 
 
 many of the sentences as possible. To construct such a 
 scale for the measurement of language ability of this 
 type was the object of the study made by Trabue. 
 
 A large number of incomplete sentences were con- 
 structed. After a preliminary trial fifty-six of these sen- 
 tences were selected and their relative difficulty deter- 
 mined by administering them, under standard conditions, 
 to several thousand children and young people in various 
 school systems. The detailed scheme by which each 
 sentence was marked will be described later, but the 
 general method was to give a score of 2 for a perfect com- 
 pletion, a score of 1 for an almost but not quite perfect 
 completion, and a score of for a failure to attempt or 
 for an imperfect completion. 
 
 By determining the different scores made on the sen- 
 tences in the various grades, it was possible to calculate 
 the relative difficulty of each of these sentences. Thus, 
 two sentences were considered of equal difficulty when 
 they were completed by the same percentage of individuals 
 tested. The greater the difference of percentage attained 
 in completing two sentences, the greater was the difference 
 in the difficulties of the sentences. It is impossible to 
 enter into the details of these calculations, but the method 
 employed was essentially the same as that described in 
 the construction of the Buckingham Spelling Scale. 
 
 Knowing the difficulty of these original sentences, 
 Trabue constructed eight short scales. The following 
 are some of the reasons for the use of several short scales : 
 
 (1) A short scale takes less time to administer and score ; 
 
 (2) a measure of ability is more reliable when taken on 
 two separate occasions than when taken at one time; 
 
 (3) a number of scales of equal difficulty admit of a class 
 being tested from time to time, the use of different scales 
 being necessary to eliminate the factor of memory. 
 
 Two scales, called by the author B and C, are here 
 shown ; in the study six similar scales are also given.
 
 Completion Test Language Scales 159 
 
 Language Scale B 
 
 Write only one word on each blank. Time limit, seven minutes. 
 
 Name 
 
 1. We like good boys girls. 
 
 6. The is barking at the cat. 
 
 8. The stars and the will shine tonight. 
 
 22. Time often more valuable money. 
 
 23. The poor baby as if it were sick. 
 
 81. She if she will. 
 
 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. 
 
 Language Scale C 
 
 Write only one word on each blank. Time limit, seven minutes. 
 
 Name 
 
 2. The sky blue. 
 
 5. Men older than boys. 
 
 12. Good boys kind their sisters. 
 
 19. The girl fell and her head. 
 
 24. The rises the morning and at night. 
 
 30. The boy who hard do well. 
 
 37. Men more to do heavy work women. 
 
 44. The sun is so that one can not 
 
 directly causing great discomfort to the eyes. 
 
 53. The knowledge of use fire is of 
 
 important things known by but unknown 
 
 animals. 
 56. One ought to great care to the right of 
 
 , for one who bad habits it 
 
 to get away from them. 
 
 The scales in this section are reproduced by the courtesy of Dr. 
 M. R. Trabue.
 
 160 Scientific Measurement 
 
 Each of these scales consists of ten steps or sentences, 
 the intervals between the various sentences being approxi- 
 mately equal ; that is, sentence 6 is as much more difficult 
 than sentence 1, as sentence 8 is more difficult than sen- 
 tence 6, and so on. It should, however, be noted that 
 Scale C is, on the whole, a little harder than Scale B, Sen- 
 tence 2 in Scale C is a little more difficult than sentence 
 1 in Scale B, and sentence 5 in Scale C is a little more 
 difficult than sentence 6 in Scale B. The same is true 
 throughout the series. 
 
 Directions for Administering the Test 
 
 The scales which have been described may be pur- 
 chased in any quantity from the Bureau of Publications, 
 Teachers College, New York. It should be noted that 
 these standard blanks must be used if the results 
 obtained are to be used for comparative purposes. 
 When the test is given to a third or lower grade, it 
 is necessary to give a little preliminary training, using a 
 practice sheet, which can be secured with the regular 
 tests. In the fourth grade and above, the following oral 
 explanation should be made before distributing any 
 papers : 
 
 This sheet contains some incomplete sentences, which form a 
 scale. This scale is to measure how carefully and rapidly you can 
 think, and especially how good you are in your language work. 
 
 You are to write one word on each blank, in each case selecting the 
 word which makes the most sensible statement. 
 
 You may have just seven minutes in which to sign your name at 
 the top of the page and write the words that are missing. The papers 
 will be passed to you with the face downward. Do not turn them 
 over until we are all ready. After the signal is given to start, re- 
 member that you are to write just one word on each blank and that 
 your score depends on the number of perfect sentences you have at 
 the end of seven minutes. 
 
 If there are no questions, the papers may then be dis- 
 tributed, taking care that no child looks at the printed
 
 Completion Test Language Scales 161 
 
 side until there is a paper upon the desk of each child 
 and the following additional instructions have been given : 
 
 After you have been working seven minutes, I shall say, "The 
 time is up. All stop writing ! " You will all please stop at once and 
 lay aside your pens (or pencils). Now if you are all ready, you may 
 turn your papers, sign your names and fill the blanks. 
 
 Take note of the exact time at which the signal to start 
 was given, allow exactly seven minutes, and give the 
 command to stop writing. Collect all papers at once. 
 It is very important that exactly seven minutes be al- 
 lowed. A stop watch is the most satisfactory means of 
 keeping the time on a test of this sort. Grade each paper 
 according to the general scheme about to be described, and 
 make a record of the total number of points made by each 
 pupil, in order that the amount of progress of each indi- 
 vidual may be determined when this scale is used for a 
 second time, or when another scale is employed. Then 
 arrange the scores in ascending order and find out the 
 median score ; namely, that point above and below which 
 there are an equal number of scores. This median value 
 may then be compared with the medians obtained by 
 other classes. 
 
 General Scheme of Scoring 
 
 The following general scheme has been the basis upon 
 which the more detailed judgments have been based : 
 
 Score 2 
 A score of 2 points is to be given each sentence completed perfectly. 
 Errors in spelling, capitalization, and punctuation should not be 
 allowed to affect the score. 
 
 Score 1 
 
 A score of 1 is to be given each sentence completed with only a 
 slight imperfection. A poorly chosen word or a common gram- 
 matical error, which makes the sentence less than perfect and yet 
 leaves it with reasonably good sense, should serve to reduce the score 
 from 2 to 1.
 
 162 Scientific Measurement 
 
 Score 
 
 A score of is to be given if the sentence as completed has its 
 sense or construction badly distorted. A sentence must have reason- 
 ably good meaning and express a sentiment which might honestly be 
 held by an intelligent person in order to receive a higher credit than 
 zero. 
 
 It is apparent that the above method of scoring leaves 
 more than is desirable to the judgment of the person 
 who is rating the sentence. This subjective element in 
 the marking is much reduced, however, by a careful con- 
 sideration of the examples given by the author of what 
 in his opinion constitutes a sentence worth the score 2, 1, 
 and 0, respectively. For illustration take the sentence : 
 
 30. The boy who hard do well. 
 
 Score 2 
 
 works, tries, studies, thinks, will, 
 
 Score 1 
 
 tries .... can, may, does, shall, should, could, must, 
 
 worked, tried, .... did, will, can, 
 
 plays, hits, work, .... will, 
 Score 
 
 tries .... sometimes, surely, often, 
 
 did ... . work did, does .... work, work .... did, 
 
 All the other sentences are treated similarly. The reader 
 is referred to the original study for these completions, as 
 they are too bulky to warrant introduction here. 
 
 It will be noticed that the score is given for the whole 
 sentence; that is, in those cases where more than one 
 blank appears, the mark is not given for each single com- 
 pletion but for the whole sentence. 
 
 To summarize : All that is necessary to test a class in 
 the type of language ability measured by these scales, 
 is to procure the standard blanks from the publishers. 
 Follow carefully the directions for the administration of 
 the test. Score the tests according to the scheme out- 
 lined. Determine the score below which and above which
 
 Completion Test Language Scales 163 
 
 there are an equal number of pupil's records, and then 
 compare this median value with previous records, if such 
 have been obtained ; if not these first results will estab- 
 lish tentative standards. 
 
 EXERCISES 
 
 1. How could you rank five completion tests, of your own con- 
 struction, according to their difficulty? What is the test of a suit- 
 able sentence for a particular group ? 
 
 2. What is the advantage of having the difficulty of the sentences 
 in the scale rise by equal increments? What would happen if three 
 of the sentences were of the same difficulty? 
 
 3. How would you use completion tests for determining whether 
 the pupils had read a certain assignment of history or geography? 
 
 4. To what extent do these completion tests measure a valuable 
 language ability? How does this type of ability compare with 
 ability in English Composition in your class ? 
 
 5. How could the idea of the completion sentence be employed to 
 measure ability in a foreign language? 
 
 6. Can you reasonably expect the same standard of work, in this 
 test, from schools in a foreign district, and in an English-speaking 
 district? How could a school of the first type establish its own 
 standards? 
 
 7. How would you determine the standard of your own class in 
 this test, as compared with other classes of the same grade? 
 
 8. State how you would compare the standing of your own school 
 with that of another school? What conditions would have to be 
 fulfilled to make this comparison justifiable? 
 
 9. Are completion sentences merely a test, or could they be used 
 with advantage as an exercise to increase thought in language lessons? 
 
 10. Suppose a grade fell notably below its average of the last few 
 years, what steps would you take to meet this decline?
 
 CHAPTER VIII 
 
 DRAWING SCALE 
 
 THORNDIKE DRAWING SCALE 
 
 The measurement of improvement and efficiency in a 
 subject such as drawing is beset with great difficulties. 
 It is reasonable to suppose that in art the judgment of 
 excellence depends on the individual teacher, to a greater 
 extent than in most of the other subjects in the school 
 course. In spite of this supposition, Thorndike in 1913 
 presented a scale which, though merely tentative, yet 
 limits to a great degree the possible differences of individual 
 opinion in estimating drawing ability. Its method of 
 derivation is very similar to that employed in the scale 
 for the measurement of English composition. From 45 
 carefully selected drawings from Kerschensteiner's "Die 
 Entwickelung der zeichnerischen Begabung," a more 
 limited selection of 14 drawings was made. These, 
 together with a drawing from another source, constituted 
 the 15 samples. These samples were then submitted to 
 artists, teachers, and students of education and psychol- 
 ogy, with the request that they be ranked in the order 
 of merit : that is, that No. 1 be assigned to the drawing 
 which, in the opinion of the judges, is the best; No. 2, 
 to the drawing that is the next best, etc. ; No. 15 being as- 
 signed to the very worst drawing. It was stated quite 
 clearly that no allowance should be made for the apparent 
 age or training of those who had made the drawings, but 
 that the drawings should all be judged by the standard 
 of their intrinsic merit. In all, 376 ratings or rankings of 
 the 15 drawings were obtained, 60 of which were from 
 
 164
 
 Drawing Scale 
 
 165 
 
 A Scale for the Merit of Drawings by 
 Pupils 8 to 15 Years Old 
 
 The numbers give the merit of the drawing as judged by 400 artists, 
 teachers of drawing and men expert in education in general
 
 166 
 
 Scientific Measurement
 
 Drawing Scale 
 
 167 
 
 00 
 
 ! 
 
 feu S \P 

 
 168 
 
 Scientific Measurement 
 
 •-$ 
 
 *<s*»^ 
 
 17.0
 
 Drawing Scale 169 
 
 artists who had sufficient merit to be included in "Who's 
 Who in America." 
 
 Suppose the drawings be called a, b, c, d, e, /, g, h, i, j, 
 k, I, m, n, and o. These fifteen drawings were so chosen 
 that they proceeded step by step from a drawing of almost 
 zero merit, to a drawing of such a high order, that only 
 one child out of five thousand under fifteen years of age 
 was able to produce work of that degree of excellence. 
 When the data which came from all the judges were col- 
 lected, an idea was obtained of the relative merit of each 
 of the samples of drawing. Thus, suppose it was desired 
 to compare sample b with sample a, and it was known 
 that 95% of the judges rated 6 as having more merit than 
 a, while 85% rated c as having more merit than b. From 
 general considerations, it was safe to assume that the dif- 
 ference in quality between sample 6 and sample a was 
 greater than the difference in quality between sample b 
 and sample c. This can be seen at once if we consider 
 what it means when 100% of the judges rank one sample 
 greater than another sample ; for, in this case, the superior 
 sample is so much better than the inferior sample, that 
 not one judge in a hundred thinks it inferior. Thus, if 
 we compare the plays of Marlowe and Shakespeare by 
 this method, there is such a great difference in quality, 
 that 100% of competent judges would think Shakespeare 
 superior to Marlowe. 
 
 Let us consider for one moment what is implied by the 
 statement that 50% of the judges ranked specimen X as 
 better than specimen Y; in this case, as many judges 
 thought Y was better than X as thought X was better 
 than 7. Under these conditions, if the judges are com- 
 petent and sufficiently numerous, we are justified in assum- 
 ing that X is equal to Y in merit. Thus, if 100 people 
 were to compare the merits of two novels, such as "Silas 
 Marner" and "Scenes from Clerical Life," and it was 
 found that 50% of the judges thought "Silas Marner"
 
 170 
 
 Scientific Measurement 
 
 was superior and 50% thought "Scenes from Clerical 
 Life" was superior, we should be justified in assuming 
 that the two novels were of approximately equal merit. 
 To summarize, when 100% of the judgments rank X as 
 superior to Y, then X is in all probability very far re- 
 moved in merit from Y; whereas, when 50% of the 
 judgments rank X as superior to Y, then X and Y are 
 approximately equal in merit. The results of the rating 
 of the drawings by 187 judges are shown below in the 
 table. 
 
 RATINGS OF DRAWINGS 
 94.85% of the judges rated b as better than a. 
 
 84.5 
 
 88.45 
 
 69.5 
 
 82.55 
 
 69.7 
 
 89.4 
 
 81.75 
 
 70. 
 
 73.35 
 
 72.5 
 
 86.5 
 
 74.2 
 
 It tt ti 
 
 " C 
 
 a 
 
 
 it it ti 
 
 " d 
 
 (( 
 
 
 a a a 
 
 " e 
 
 a 
 
 
 a it ti 
 
 " f 
 
 a 
 
 
 tt tt it 
 
 " 9 
 
 a 
 
 
 ft a tt 
 
 " h 
 
 St 
 
 
 tt it it 
 
 " i 
 
 a 
 
 
 it it it 
 
 " J 
 
 a 
 
 
 a a it 
 
 " k 
 
 a 
 
 
 a a tt 
 
 " I 
 
 a 
 
 
 n tt tt 
 
 " m 
 
 tt 
 
 
 n tt a 
 
 " n 
 
 a 
 
 a 
 
 b. 
 c. 
 
 d. 
 e. 
 f. 
 9- 
 h. 
 i. 
 
 h 
 
 k. 
 
 I. 
 m. 
 
 By simple, but laborious statistical treatment, which it 
 is unnecessary to discuss here, based on the two facts 
 given above, it is possible to arrange the various samples 
 in an order of merit, and to assign to each a numerical 
 value which is the result not of a single judgment but of 
 the combined estimates of many experts. That is, if .we 
 assign zero merit to the first picture, which is supposed 
 to be a picture of a man, then by an analysis of the 
 table just given it can be shown that the second figure, 
 which is intended to be a house, has 2.4 degrees of 
 merit. On the same scale, Figure 3, which is also 
 supposed to be a house, has 3.9 degrees of merit, and
 
 Drawing Scale 171 
 
 so on, until we reach the last three samples, which 
 have 14.4, 16, and 17 degrees of merit, respectively. A 
 scale so constructed enables us to measure skill and im- 
 provement in drawing by methods which are largely 
 objective. 
 
 In the matter of assigning actual values to the drawings, 
 care must be taken not to assume that the degree of im- 
 provement, say from a sample which ranks 6 to a sample 
 ranking 10, is equal to that from a sample ranking 12 to 
 one ranking 16, in the sense that a rise in a temperature 
 scale from 6° to 10° is equal to a rise from 12° to 16°. 
 In the case of the scale for measuring drawing, this is 
 true in a very limited sense only, but the scale can be 
 used with a maximum return without an understanding 
 of these statistical considerations. In other words, when 
 a teacher says that the average ability of a class accord- 
 ing to the Thorndike Drawing Scale is 13.5, it conveys a 
 reasonably definite idea to any other person who is ac- 
 quainted with that scale. For all practical purposes the 
 samples constituting the scale, as used by the average 
 teacher, might have been lettered instead of having nu- 
 merical values attached. 
 
 When it is desired to measure the ability of a class by 
 using this scale, all that is necessary is to choose a certain 
 model or subject and allow a measured time for the 
 drawing; this time should be varied according to the 
 nature of the subject. The subject and the time allowed 
 should be noted very carefully, so that when the test is 
 given again, all these external conditions may be the 
 same. When the drawings are collected each one is 
 measured by being placed alongside the scale, and its 
 position estimated by the teacher, or by several teachers. 
 If it appears to lie between two points of the scale, an 
 intermediate value may be given. 
 
 This still leaves a considerable amount to the individual 
 judgment. In other words, the scale is not by any means
 
 172 Scientific Measurement 
 
 purely objective, for equally competent persons would 
 fail to assign the same degree of merit to the same draw- 
 ing. This factor of personal opinion can however be 
 curtailed by having several individuals measure the 
 drawing by the scale and taking the average of their 
 judgments. The drawing scale, like any other scale in 
 its beginnings, is very incomplete, since it still remains 
 to work out scales for all the various types of drawings 
 taught in the schools. But any scale is better than no 
 scale at all, and continued use of a drawing scale by 
 teachers will standardize judgments and encourage quan- 
 titative thinking, even in this study which at present is 
 so dependent on personal opinion. 
 
 EXERCISES 
 
 1. Take 30 specimens of drawing, distributed through the grades, 
 and mark them according to your usual method ; let one month 
 elapse and grade them again. What do the results show? 
 
 2. Repeat the above experiment, with the exception that the 
 grading is done by means of the Thorndike Scale. How do the two 
 ratings differ? Compare with the results of the previous experiment. 
 
 3. Why is it necessary to take careful note of the time allowed 
 for the test? Why must this be the same when the test is repeated, 
 if the grading is to be used to measure improvement? 
 
 4. Why cannot we divide children into two classes — "good 
 drawers" and "bad drawers"? 
 
 5. How would you proceed to establish norms or standards for 
 drawing ability in the various grades in your school? 
 
 6. On the analogy of the Harvard-Newton Scale, how would you 
 propose to construct a better method of measuring drawing ability? 
 
 7. Take 5 specimens of drawing from each of the grades; have 
 these rated on the scale by 5 different individuals. How would these 
 results give you an indication of the reliability of the scale?
 
 CHAPTER IX 
 THE APPLICATION OF THE SCALES IN THE SCHOOLS 
 
 Objective Scales in Other Subjects. Scales for the 
 measurement of other school products have yet to be 
 evolved ; the subject is still in its beginnings. In addi- 
 tion to those scales already described, attempts have been 
 made to measure objectively mechanical constructive 
 ability and ability in the translation of Latin, while scales 
 are in process of formation for the measurement of ability 
 in several of the modern languages, in algebra and geom- 
 etry, and in some of the natural sciences. One of the 
 authors is at present conducting an experiment, extending 
 over two years, the results of which will standardize the 
 rate of improvement in typewriting using the touch 
 method. 
 
 A point of interest arises as to whether scales can be 
 worked out for informational subjects such as history, 
 geography, etc. ; for at once we have to face the great 
 difficulty that in subjects such as these we have to meas- 
 ure knowledge of facts or content rather than skill or 
 method. The previous study of the writing scale has 
 little effect on a child's proficiency in writing, but the 
 study of a scale for the measurement of content or facts 
 in history, prior to the examination, renders that scale 
 valueless as a test. For the particular facts can be 
 learned, and the knowledge of these will not indicate any 
 general knowledge of the whole field. This means that 
 in measuring efficiency in certain subjects, we may have 
 to resort to analysis, and use one objective standardized 
 test to measure method, and another more or less sub- 
 
 173
 
 174 Scientific Measurement 
 
 jective test to measure content. Whether it will ever be 
 possible to use a universal and unchanging scale for 
 content values remains doubtful. If a very large num- 
 ber of content questions, sufficiently wide to cover the 
 field, could be standardized as regards difficulty, there is 
 no reason why a purely objective scale, consisting of a 
 few of these questions selected at random, should not be 
 employed. 
 
 The adoption of these objective scales for the measure- 
 ment of school products is bound to establish a scientific 
 attitude in the schools, which will energize and direct the 
 work of the teachers and raise the administrator's task 
 from the realm of mere opinion to the level of scientific 
 judgment. 
 
 Standardization of the Objective Scales. The standard- 
 ization of these universal tests will involve a considerable 
 amount of work if accurate and complete norms are 
 to be established. In some cases it may be advisable to 
 have the test standardized not only as regards the prod- 
 uct of each grade, but also with reference to age. For 
 example, in handwriting — a distinctly motor function — 
 it may be well to know the quality of work expected at a 
 certain age, as well as in a certain grade. A pupil may 
 be held back in a grade because of failure in arithmetic 
 and reading and so become over age. Under these condi- 
 tions a motor function such as handwriting may continue 
 to improve normally, so that even though the child is in 
 a low grade, we may expect of him the normal standard 
 product of his age in that subject. 
 
 As these tests come nto common use in the classroom, 
 the interest of the individual teacher will cease to be con- 
 fined merely to the average of the particular grade, and 
 attention will more and more be directed to deviations 
 from that average which may normally be expected. In 
 fact one of the great services of these tests is to reveal 
 the great individual differences in ability that exist even
 
 The Application of the Scales in the Schools 175 
 
 in the same class. A teacher of Grade V will not only 
 be interested in knowing that the average achievement of 
 the class in a particular test, say in the Courtis Test, 
 Fundamentals 7, should be 9.0, but will also appreciate 
 the advantage of knowing how the class groups itself 
 around this average, what are the extreme deviations in- 
 dicating the lowest and highest type of work in the class. 
 In fact, as will be seen, these scales may be used for a 
 variety of purposes by a teacher who is really interested 
 in the work of the individuals of the class. 
 
 The Relation of the Objective Scales to Continuous School 
 Records. The application of statistics to education is 
 not really a new idea; in certain realms of adminis- 
 tration, such as attendance, per capita cost, etc., such 
 measurements have always been made. What is claimed 
 is that this method should extend to all possible phases 
 of classroom work. The ordinary examination fails to do 
 this; the questions are arbitrary, they are not weighted 
 according to their relative values, there are no objective 
 standards of accomplishment. Until units of mass, 
 length, and time were invented upon which all agreed, it 
 was impossible to express weight, dimensions, and time 
 in terms which would convey the same meaning to all. 
 This was, to a great extent, the situation in education 
 ten years ago. But the time is not far distant when, in 
 many of the essential subjects, the progress of every 
 pupil who enters school will be determined by objective 
 methods. Thus, in a particular function such as writ- 
 ing, we shall measure the ability of the pupil every six 
 months from the time he enters until he leaves school. 
 The same will be true of his ability in the other subjects 
 which the school considers to be of importance and which 
 admit of being measured by universal standards. The 
 enlightened school system will have the progress of every 
 child kept on a chart, a rough sample of which is given 
 on the following page.
 
 176 
 
 Scientific Measurement 
 
 m 
 H 
 
 Eh 
 H 
 
 > 
 
 H 
 
 i— i 
 
 w 
 
 o 
 
 I— I 
 
 Pn I 
 o 
 
 w 
 
 u 
 
 Q 
 P3 
 O 
 o 
 
 Pn 
 
 s 
 
 H 
 O 
 
 o 
 
 o 
 o 
 x 
 o 
 
 00 
 
 r-t OJ CO -rj< ® 
 
 
 IS 
 
 IS 
 
 (3 
 a 
 
 1-3 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 11 
 
 i-l 
 
 •4 
 
 01 
 
 l-s 
 
 
 
 
 
 
 
 
 
 
 
 
 a 
 
 
 
 
 
 
 
 
 
 
 
 
 N 
 O 
 
 Ha 
 
 
 
 
 
 
 
 
 
 
 
 
 a 
 
 OS 
 
 l-s 
 
 
 
 
 
 
 
 
 
 
 
 
 M 
 
 o 
 a 
 3 
 
 1-3 
 
 
 
 
 
 
 
 
 
 
 
 
 a 
 
 >-3 
 
 
 
 
 
 
 
 
 
 
 
 
 
 a 
 a 
 
 1-3 
 
 
 
 
 
 
 
 
 
 
 
 
 a 
 
 ai 
 
 
 
 
 
 
 
 
 
 
 
 
 O 
 
 a 
 
 a 
 
 >-3 
 
 
 
 
 
 
 
 
 
 
 
 
 a 
 
 d 
 
 1-3 
 
 
 
 
 
 
 
 
 
 
 
 
 H 
 O 
 
 a 
 
 n 
 
 & 
 
 bo 
 
 a 
 
 c 
 
 T-l 
 
 m 
 
 H 
 
 '-13 
 
 a 
 
 rC 
 
 -t-> 
 
 < 
 
 4-1 
 
 o 
 •»— i 
 
 0) 
 
 s 
 
 co 
 
 +i 
 
 a: 
 ai 
 Eh 
 
 « 
 
 •i— » 
 
 0) 
 
 a 
 
 < 
 
 o> 
 
 <* 
 
 +j 
 
 CQ 
 <D 
 
 Eh 
 o 
 
 — < 
 
 a 
 3 
 
 13 
 >-. 
 
 bo 
 
 q 
 
 • i— < 
 
 a 
 
 c 
 
 •t— » 
 
 00 
 
 bo 
 C 
 
 a) 
 
 'Lo 
 o 
 
 a 
 
 S 
 o 
 O 
 
 co 
 
 "Ho 
 fl 
 W 
 
 co 
 a> 
 H 
 
 a; 
 bo 
 as 
 3 
 bo 
 
 as 
 h3 
 
 bo 
 C 
 
 "a; 
 ft 
 in 
 
 bo 
 
 S3 
 
 Q
 
 The Application of the Scales in the Schools 177 
 
 In this way it will be a simple matter to determine the 
 exact point at which the pupil failed to advance at the 
 normal rate in any particular line of study. The teacher 
 will be able to see whether failure was confined to a par- 
 ticular subject or whether it also took place in other sub- 
 jects in the curriculum. If it is found that the child has 
 failed in but one subject and not in the others, then we 
 must assume either that the child was abnormal in that 
 subject or that the method of instruction in that particu- 
 lar branch was not up to the usual standard. If, in addi- 
 tion to this, it is found that the majority of pupils under 
 a particular teacher have failed to advance in this 
 subject alone and not in others, then there is every reason 
 to suppose that it was not the fault of the class but rather 
 the fault of the teacher in failing to give attention to the 
 subject or in using some method which could not produce 
 the average rate of progress. Again, in the case of a 
 particular pupil, it may be found that the failure to 
 progress was not confined to one subject, but that it 
 extended to all subjects. In this case further inquiries 
 must be made. It may have been a matter of arrested 
 mental development, or it may have been due to physical 
 causes or to social conditions which did not admit of the 
 child's spending sufficient time in school. 
 
 Such a chart as we have shown can easily be passed on 
 from school to school as the child goes from one neigh- 
 borhood to another, or it can be passed from school sys- 
 tem to school system or from country to country, for the 
 very essence of these universal scales is that they are 
 independent of place and time. School systems, under 
 these conditions, will keep track of every pupil from the 
 time he enters to the time he leaves. In other words, 
 the administrator will cease to deal with mere groups of 
 children and will deal with the individual child. 
 
 Application of the Objective Scales to the Question of 
 Promotion. Such methods of measurement will bring to
 
 178 Scientific Measurement 
 
 the question of promotion a definiteness which is sorely 
 needed. It is too well known that in many school sys- 
 tems a high percentage promotion does not mean a high 
 standard of work, but rather a lowering of that standard 
 to enable the requisite number of children to pass. As a 
 result, pupils are often found in the higher grades who are 
 totally unable to profit by the relatively advanced in- 
 struction given. As long as the present loose methods 
 of measuring school achievements are in vogue, such a 
 state of affairs is inevitable; under the new system a 
 radical change will be possible, for with certain exceptions 
 the presence of a child in a particular grade must mean 
 that he has passed certain points on the scales which 
 measure the various school abilities. If these points have 
 not been reached, then the pupil will not be promoted, 
 for he will be unable to profit by the instruction given. 
 
 A teacher will be able to measure the abilities of pupils 
 when they are received in September, and if promotion 
 has taken place in spite of bad previous records, he will 
 at least know of this, and, by pointing to their records, 
 will be able to free himself from criticism on account of 
 their ultimate failure. The position of the efficient and 
 conscientious teacher will be established, not on the 
 insecure basis of the opinion of an often prejudiced super- 
 visor, but on the basis of the actual work of the pupils 
 judged by impartial standards. 
 
 Application of Objective Scales to Vocational Guidance. 
 Such a chart of improvement will be of great service 
 when the pupil on leaving school requires vocational 
 guidance. The employer will state the requirements of 
 his work in the different school subjects, while the voca- 
 tional guidance expert, by consulting the chart, can deter- 
 mine the extent to which the pupil measures up to these 
 requirements. 
 
 The Objective Scale as Limiting the Amount of Improve- 
 ment Necessary. Again, it is true that in many subjects
 
 The Application of the Scales in the Schools 179 
 
 only a certain degree of efficiency is demanded by the 
 world. For example, there is no object in being able to 
 write better than is required for reasonable grace and 
 legibility. The handwriting of some children shows a 
 wasted youth ! If time is spent beyond a certain point, 
 it is relatively wasted. Yet what guarantee have we 
 that when children reach this point they will no longer 
 be given writing lessons? Under the present subjective 
 system of measurement such a guarantee is impossible, 
 and, if given, is meaningless. When the objective scale 
 is used for measuring handwriting, the matter is perfectly 
 simple ; for the child knows that when he reaches a cer- 
 tain point on the scale, provided he keeps up to that 
 point, all formal writing lessons will cease. 
 
 Application of the Objective Scales in Rural Schools. 
 These scales will find ready application in the rural schools, 
 where the teacher is unable to form correct estimates of 
 the work because small classes do not afford a basis for 
 judgment. With the new methods which these scales 
 introduce, the isolated child in the rural school can be 
 compared with, and in a sense can compete with, children 
 of like age in the city system. In fact, at present one of 
 the authors is comparing, by means of these universal 
 standards, the work of 100 rural school children of a 
 given age, with a random sampling of 100 city school 
 children of the same age. In a sense the results of this 
 experiment will be as definite as measurements made of 
 the pupils' height and weight by means of the foot-rule 
 and the weighing machine. 
 
 The Scales as Revealing the Success and Failure of School 
 Methods. The purpose of these scales, in fact of the 
 whole subject of educational measurements, is not, like 
 the ordinary examination, to test merely the efficiency of 
 the individual teacher or pupil, but rather to test the effi- 
 ciency of the teaching process itself. The individuals 
 are examined in many cases, not because of our interest
 
 180 Scientific Measurement 
 
 in them as individuals, but because their work will reveal 
 whether the method which is being used in their instruc- 
 tion is sound. Many of the failures in our schools are 
 due, not to unavoidable inefficiency on the part of the 
 teachers, but rather to lack of knowledge on their part 
 that their efforts are failing to produce the desired results. 
 Were the teachers themselves aware that they were fail- 
 ing, they would certainly attempt to alter their methods. 
 It is lack of definite knowledge of what the pupils are 
 accomplishing, and not incompetence or indifference, 
 which prevents a better adaptation of method to product 
 desired. 
 
 For this reason teachers should be willing and eager to 
 submit their work to an impersonal standard, not so that 
 it may be praised or condemned, but so that they them- 
 selves may know whether their methods are producing 
 as good results as may reasonably be expected. Teachers 
 should have a more exact knowledge than they have had 
 in the past of those processes which are going on in their 
 pupils, for it is the changes which occur during the school 
 period that must be measured. Over these changes we 
 have more or less direct control ; the test of life is too re- 
 mote. The application of these objective scales enables 
 the teacher to know what is happening, not in terms of 
 mere empty formulae which unfortunately have become 
 associated with the word "method," but rather in terms 
 of what the pupils can actually do as a result of the in- 
 struction given them. 
 
 Scientific measurement in education will narrow the 
 limits of the wasteful trial and error method which is 
 always incident to the teaching process, however con- 
 scientious the teachers may be. It will also do another 
 great service, for it is undeniable that, by means of these 
 scales, the complacency of a small section of teachers can 
 be disturbed by actually showing them their failure in 
 black and white. The greatest check on inefficiency in
 
 The Application of the Scales in the Schools 181 
 
 any system is the knowledge that the work of each teacher 
 and the work of each school can be compared with the 
 work of other teachers and the work of other schools. A 
 school which is confronted with indisputable evidence of 
 its shortcomings is in a position to investigate causes, and 
 if necessary to trace them to individuals ; such procedure 
 is always the forerunner of progress. 
 
 EXERCISES 
 
 1. What would be the chief difficulties in constructing a scale for 
 the measurement of knowledge of American history in the eighth 
 grade? 
 
 2. How would you prove to an outsider that there v are great in- 
 dividual differences in ability, even in the same class? How should 
 a knowledge of these individual differences affect (a) the amount of 
 matter taught; (b) the method of instruction? 
 
 3. What are the chief advantages of continuous school records? 
 Draw up a table, and outline the methods which could be used for 
 recording a child's progress in the fundamental studies, from the time 
 he enters to the time he leaves school. 
 
 4. Upon what factors should promotion depend? Have we any 
 right to promote a pupil if he is not up to certain minimum stand- 
 ards ? How do the standard scales help to determine these minimum 
 standards? How does too lax a promotion system disorganize the 
 work in the higher grades? 
 
 5. Why is the present system of marking in your school an 
 insufficient guide to the quality of the work which is being done? 
 
 6. Should all children give the same time to all studies? In what 
 way will the use of these standard tests enable us to allow the indi- 
 vidual child to distribute his time in a more advantageous manner? 
 
 7. How is a rural school teacher handicapped in judging the work 
 of her pupils? Show how the scales help in this respect. 
 
 8. A superintendent of a city school system cannot decide between 
 two proposed methods of teaching handwriting. Describe a plan 
 whereby, in a few years, he could decide which method was the 
 better? How have such questions been decided in the past? 
 
 9. Why is it better to measure the success of a year's work by 
 the improvement of the pupils during that period, than by the final 
 scores in a test at the end of the year? If you were the principal of a 
 school, outline the methods you would employ to measure such 
 improvement.
 
 CHAPTER X 
 
 DANGERS INCIDENTAL TO THE USE OF THESE 
 
 SCALES 
 
 At a time when all available pressure should be brought 
 to bear on school systems to introduce objective measure- 
 ment into the ordinary routine of the school, it seems 
 hardly the occasion to criticise the scales. However, a 
 word of caution may not be out of place as to the dangers 
 which may arise from their application, since their im- 
 proper use will perhaps prejudice those who make the 
 first attempts at this type of measurement. 
 
 Difficulty of Comparing Methods of Teaching. It has 
 already been stated that one of the great functions of the 
 scales is to compare the various methods of instruction 
 employed in the teaching of a subject. Great care, how- 
 ever, will have to be taken to prevent mistakes in com- 
 paring the relative values of such methods when used in 
 different schools or systems. To know that the work in 
 a particular subject is better in one school than in another 
 is not sufficient to justify the judgment that the method 
 used in the one school is superior to that in the other. 
 In such a comparison several secondary causes must also 
 be considered before any statement is made concerning 
 the relative efficiency of the methods : (1) time allowed 
 in the different schools; (2) personality of the teacher; 
 (3) the type of neighborhood as determining the type of 
 pupil. It will be only by the most careful experimenta- 
 tion, where attention is paid to these points and to many 
 others of less importance, that anything like a scientific 
 application of the scales to the question of the values of 
 
 182
 
 Dangers Incidental to Use of These Scales 183 
 
 methods will be obtained. The whole subject is full of 
 danger, and many fallacies will have to be avoided. At 
 the present time scientific attention is being directed 
 rather to the construction and use of scales for particular 
 groups than to comparison of procedure values ; but such 
 comparison will be possible later, when every school sys- 
 tem employs a competent statistician and experimenter 
 capable of conducting genuinely scientific comparative 
 experiments. In short, we must not strive to compare 
 groups that are not alike or hold up standards without 
 due consideration of social conditions. Mere statistics 
 can never dictate final standards of achievement ; a 
 standard set up may be too high for one school and not 
 high enough for another. Each school, after working 
 with these scales for some time, can establish standards 
 of its own ; but there is always the danger that a standard 
 may be set up which falls short of what should be done. 
 In fact, the unwise use of standards, in this respect, may 
 confirm the school in lax processes. 
 
 Failure of Scales, from the Fact That They Measure Com- 
 plex Abilities, to Reveal the Point of Weakness in Method. 
 While these scales will do much to quicken methods used 
 in the schools, it may be well to mention another point 
 which is apt to be overlooked by some who employ such 
 measurements. Thus, a scale may show that the method 
 which has been used is imperfect in that it has failed to 
 produce the desired product ; but it does not directly 
 analyze the particular fault. The scales do not tell you 
 what to do, but rather they tell you where you are. A 
 teacher may be conscious that he has failed, but unable, 
 in spite of great efforts, to find out the exact factor re- 
 sponsible for this failure. In much the same way a phy- 
 sician after examination may make the announcement 
 that the organic processes are wrong, but at the same 
 time be totally unable to attribute the cause. Although 
 the present scales, because they measure such complex
 
 184 Scientific Measurement 
 
 activities, do not reveal the exact point at which a teacher 
 may have failed, yet we see in the Courtis Test the begin- 
 nings of an attempt to measure the details of what many 
 have considered to be a single process, namely, "arith- 
 metic ability." When more analytical scales have been 
 worked out in other subjects, it will be possible to go into 
 detail and tell the teacher at just what point or points he 
 failed, these small failures accounting for the failure in 
 the wider test. The idea might also be applied to the 
 testing of English composition. As things are now, it is 
 possible merely to tell a teacher that the class has failed 
 to produce as good English composition, as measured on 
 the Hillegas Scale, as might be expected. We are not in 
 a position to say what details are responsible for the 
 failure. But suppose at a later time scales should be 
 used to test (1) punctuation, (2) extent of vocabulary, 
 (3) choice of vocabulary, (4) power of summarization, 
 etc. ; then that which we now attribute perforce to general 
 weakness, we shall then assign to weakness in one or 
 more of these factors which can be corrected by special 
 practice. In this way we shall narrow down the limits 
 within which the teaching process can fail without even 
 a knowledge on the part of the teacher that it is failing. 
 
 What the Scales Do Not Measure. Another objection 
 which may be urged against the scales is that they fail 
 to take into account such factors as interest in the process 
 of learning, the eagerness with which pupils will continue 
 a particular study after pressure is removed, etc. The 
 scale also takes no direct account of the method by which 
 the product is obtained ; it does not tell the experimenter 
 whether these results were secured by easy work or by 
 undue pressure on the part of the teacher. The reply is 
 that it is only the objectors who have ever assumed that 
 the scales do measure these things. To illustrate, in an 
 automobile reliability test, the measurement of speed 
 does not tell us concerning the internal mechanism of the
 
 Dangers Incidental to Use of These Scales 185 
 
 engine ; other tests must be used to measure this factor. 
 But if a machine keeps up a high speed for a long period, 
 then as a rule the internal factors cannot be much out of 
 gear. In a precisely similar manner, if a class steadily 
 keeps up its improvement on a particular scale, then it is 
 feasible to assume that the internal factors are not seri- 
 ously wrong. In the end, bad psychological methods 
 such as undue driving (which is little to be feared in 
 modern education), will yield poor objective results. 
 The scales, however, must not be attacked because they 
 fail in many cases to measure what no competent individual 
 has ever claimed they do measure. 
 
 The use of scales also brings with it the danger that the 
 teacher may sacrifice everything in the classroom to the 
 production of work which can be measured objectively, 
 and, as already pointed out, the scales may fail to give 
 sound relative values to different elements involved in 
 that work. To make this point clearer, let us consider 
 for one moment a scale for the measurement of the child's 
 ability to add simple numbers, such as was described in 
 the Courtis test. If the norms insist upon speed, then 
 the teacher will work for speed ; if the norm is one for 
 accuracy, then the teacher will work for accuracy; and 
 the scale itself does not decide to which of these two fac- 
 tors the greater attention should be given. Even when 
 the scale is placed in the hands of the teacher, these ques- 
 tions of relative value must still be decided. However, 
 in this particular respect the scales themselves will work 
 out their own salvation, for, by a consensus of expert 
 opinion, it will be possible to decide for any particular 
 grade the amount of speed that should be required as 
 well as the degree of accuracy. 
 
 Another point against which school systems must care- 
 fully guard themselves, when these scales and standards 
 are introduced, will be a tendency for schools to overlook 
 those factors which do not admit of measurement by
 
 186 Scientific Measurement 
 
 such objective scales. This danger will gradually be 
 eliminated as time goes on and as further scales for the 
 measurement of school products are worked out. In the 
 meantime, merely because only certain abilities at present 
 admit of measurement, the school must not overlook sub- 
 jects and factors which as yet do not admit of such quan- 
 titative estimation. In particular it must not fail to 
 take into consideration such factors as the personal char- 
 acter of the teacher, the moral atmosphere of the school, 
 and other spiritual values which, like life, beauty and 
 happiness, are, to say the least, difficult fields for quanti- 
 tative analysis. Such spiritual values in schools are of 
 the greatest importance; to overlook or underestimate 
 this fact would indicate a profound lack of sense of rela- 
 tive values. Even statisticians remember these things. 
 But because we cannot estimate spiritual values, it is no 
 reason why we should not measure values in those realms 
 which admit of measurement. No science would have 
 evolved, if it had not in its beginning confined itself to a 
 limited field, and left large parts of the subject for the 
 future. Furthermore, there is very strong a priori evi- 
 dence to suggest that there is a close correlation existing 
 between spiritual values and the values which these scales 
 measure. If in the things we can measure it can be shown 
 that the work is inadequate, there is every reason to 
 believe that in the region of spiritual values there are 
 shortcomings which escape our measuring rod. Cer- 
 tainly low objective values are no great argument for high 
 spiritual values ! 
 
 The Future of Educational Measurement. Many 'of 
 these tests need criticism and revision, and such questions 
 as their fairness and practicability can be answered only by 
 the teachers who use them. For this reason the authors 
 have refrained from any detailed consideration of the 
 shortcomings of the individual scales. But the time spent 
 upon their application will accomplish a twofold purpose :
 
 Dangers Incidental to Use of These Scales 187 
 
 It will improve the scales themselves ; and it will give to 
 every teacher who employs them a quantitative point of 
 view which is sadly lacking in the schools, for many 
 questions of school procedure do not admit of being 
 answered by a mere affirmative or negative — the answer 
 is found in the quantitative measurement. The Director 
 of Reference and Research of the Department of Educa- 
 tion of the City of New York says: "There could be no 
 better exercise for a teachers' seminar than a series of 
 discussions on some selected tests that would invite the 
 independent judgment and criticism of intelligent 
 teachers." 
 
 It is dangerous to forecast, especially when a subject is 
 in its infancy, but there is every reason to believe that 
 the application of the scientific method and the logic of 
 statistics to educational problems will slowly revolution- 
 ize the method of education, even on its philosophical 
 side. Moreover, in certain branches it will raise the 
 study of education to the level of an exact science, 
 thereby winning the respect of the scientific world for a 
 subject whose low standards of proof and loose methods 
 in the past have been responsible for the stigma which 
 attaches to the study of education as an academic subject 
 in the school and college curriculum. 
 
 EXERCISES 
 
 1. When we are told that a child is "poor" in arithmetic, what is 
 implied by this statement? How may we use the scales described 
 to discover the point at which, and the extent to which, the individual 
 is below standard? 
 
 2. How may the norms established for the scales actually confirm 
 a school in lax teaching methods? How could this evil be prevented? 
 
 3. What other scales would be useful in the classroom? 
 
 4. How would you start to construct a rough objective scale for 
 measuring (a) moral judgment, (6) aesthetic appreciation, and (c) 
 humor?
 
 188 Scientific Measurement 
 
 5. How will the norms established by the use of these scales help 
 greatly in settling the question of time distribution in the schedule ? 
 
 6. Why is a single survey of a school of limited value ? What are 
 the advantages of measuring the quality of the work every half 
 year? 
 
 7. How would you show a class the rate at which it was improv- 
 ing, from month to month, in order to accelerate its progress in 
 (a) spelling, (6) writing, (c) reading? 
 
 8. How would you proceed to compare two different methods of 
 teaching spelling by means of the objective scales? Enumerate the 
 dangers and show how you would avoid them? 
 
 9. It is sometimes said, "These scales do not measure the most 
 important work of the school, therefore they are of little avail." 
 How would you meet this criticism? 
 
 10. How would you conduct, in a small city system, a general 
 survey of the quality of the work done in the common subjects of the 
 curriculum ?
 
 APPENDIX 
 
 SOURCES OF THE SCALES 
 
 The sources, from which a full account of each of the 
 scales can be obtained, are given below. 
 
 Courtis, S. A. 
 
 A Manual of Instructions for Giving and Scoring the 
 Courtis Standard Tests. (75 cents.) 82 Eliot Street, 
 Detroit. 
 
 This manual also includes the Courtis Handwriting and Reading 
 
 Scales. 
 
 The standard blanks for any of the above tests, together with full 
 directions for administration and scoring of the test, may be obtained 
 from Mr. S. A. Courtis at the above address. 
 
 Thorndike, E. L. 
 
 Handwriting. Teachers College Record, 11 : No. 2. 
 
 1910. (30 cents.) Publication Bureau, Teachers 
 
 College, New York City. 
 Separate copies of the scale can also be secured (5 cents). 
 
 Ayres, L. P. 
 
 A Scale for Measuring the Quality of Handwriting of 
 School Children. (5 cents.) Russell Sage Founda- 
 tion, New York City. 
 
 Thorndike, E. L. 
 
 The Measurement of Ability in Reading. Teachers 
 College Record, 15 : No. 4, 1914. (30 cents.) Pub- 
 lication Bureau, Teachers College, New York City. 
 
 The standard blanks used in the Thorndike Tests may be procured 
 in any quantity from the above address. 
 
 189
 
 / 
 
 190 Appendix 
 
 Starch, D. 
 
 The Measurement of Efficiency in Reading. Journal of 
 Educational Psychology, January, 1915. (30 cents.) 
 Warwick and York, Inc., Baltimore. 
 
 The standard blanks for the administration of the test may be 
 obtained, in any quantity, from the author, Dr. Daniel Starch, Uni- 
 versity of Wisconsin. 
 
 Buckingham, B. R. 
 
 Spelling Ability: Its Measurement and Distribution. 
 (95 cents.) Publication Bureau, Teachers College, 
 New York City. 
 
 Starch, D. 
 
 The Measurement of Efficiency in Spelling. Journal of 
 Educational Psychology, March, 1915. (30 cents.) 
 Warwick and York, Inc., Baltimore. 
 
 Ayres, L. P. 
 
 A Measuring Scale for Ability in Spelling. (30 cents.) 
 Russell Sage Foundation, New York City. 
 
 HlLLEGAS, M. B. 
 
 A Scale for the Measurement of Quality in English Com- 
 position by Young People. Teachers College Record, 
 13 : No. 4. 1912. (30 cents.) Publication Bureau, 
 Teachers College, New York City. 
 
 Ballou, F. W. 
 
 Scales for the Measurement of English Composition. 
 (40 cents.) The University Press, Harvard Univer- 
 sity, Cambridge, Mass. 
 
 Trabue, M. R. 
 
 Completion Test Language Scales. ($1.15.) Publica- 
 tion Bureau, Teachers College, New York City. 
 
 The scales described, together with the Practice Sheet, may be 
 purchased in any quantity from the above address.
 
 Appendix 191 
 
 Thorndike, E. L. 
 
 The Measurement of Achievement in Drawing. Teachers 
 College Record, 14: No. 5. 1913. (30 cents.) 
 Publication Bureau, Teachers College, New York City. 
 
 Woody, C. 
 
 Measurements of Some Achievements in Arithmetic. 
 (95 cents.) Publication Bureau, Teachers College, 
 New York City. 
 
 The standard blanks for the administration of these tests may 
 be procured in any quantity from the above address. 
 
 BOOKS FOR FURTHER REFERENCE 
 General 
 
 Starch, D. 
 
 Educational Measurements. The Macmillan Company. 
 ($1.25.) 
 
 Teachers Year Book of Educational References. Pub- 
 lications No. 6 and No. 14. Department of Educa- 
 tion, City of New York. 
 
 Both the above books give very adequate bibliographies. 
 
 Application of Scientific Measurement to a School Survey 
 
 JUDD, C. H. 
 
 Measuring the Work of the Public Schools. (50 cents.) 
 Survey Committee of the Cleveland Foundation, 
 Cleveland, Ohio. 
 
 17
 
 AA 000 714 945 3 
 
 •