LB 
 
 i? 
 

 IP 
 
 m i 
 
 
 
 n 
 
THE HARVARD-NEWTON BULLETINS 
 
 NUMBER II 
 
 * SCALES FOR THE 
 
 MEASUREMENT OF ENGLISH 
 COMPOSITIONS 
 
 BY 
 
 FRANK WASHINGTON BALLOU, PH.D. 
 
 DIRECTOR, DEPARTMENT OP EDUCATIONAL INVESTIGATION 
 AND MEASUREMENT, BOSTON PUBLIC SCHOOLS 
 
 CAMBRIDGE, MASS. 
 PUBLISHED BY HARVARD UNIVERSITY 
 
 SEPTEMBER, 1914 
 
(b 
 
INTRODUCTION 
 
 WHEN educational research in a city school system is carried 
 on jointly by a department of education in a university and the 
 school authorities of that city, it combines two purposes. The 
 department of education engages in it to promote the scientific 
 study of educational problems and to assist in their solution. Its 
 purpose is larger than that of the school system in which the work 
 is done. The city school system engages in such work chiefly 
 to improve educational conditions and practices in that city. 
 Such educational research as is undertaken, therefore, should be 
 of mutual advantage and should combine both purposes. 
 
 The method of carrying on such research should be coopera- 
 tive. There should be combined and applied to the solution of 
 such educational problems as are studied, the practical knowl- 
 edge which teachers, principals and other school workers possess, 
 and the educational insight and exact methods of investigation 
 which the department of education has at its command. To be 
 of the greatest value to the school system such work should be 
 participated in by as large a number of teachers and school 
 officers as can advantageously be used: and the work should 
 never be allowed to become a burden to them. On the other 
 hand, the work should be in charge of a responsible head con- 
 nected with the department of education who should apply the 
 available resources of the department to the study of problems 
 in the school system. Properly qualified graduate students can 
 participate with advantage and profit to the school system and to 
 themselves. 
 
 In the prosecution of their research, much purely clerical work 
 is inevitable. This should not be done by the teachers or those 
 in the department of education, but by competent clerks. 
 
 This study was started and has been carried on in accordance 
 with the foregoing principles. The study has been directed by 
 
4 ; INf REDUCTION 
 
 the Joseph Lee Fellow for Research in Education: it has been 
 participated in by principals and teachers in Newton, and by 
 properly qualified graduate students in the Division of Educa- 
 tion, Harvard University. At all times, the work has had the 
 careful supervision of members of the Division of Education, and 
 of Superintendent F. E. Spaulding of Newton. 
 
 At the close of a piece of cooperative research, the director 
 always finds himself under many obligations. To those who have 
 participated in this undertaking, and to those who have super- 
 vised the carrying on of the work, the writer acknowledges his 
 indebtedness, and wishes to express his sincere appreciation of 
 their hearty and effective assistance. 
 
 FRANK W. BALLOU. 
 
 
THE 
 HARVARD-NEWTON BULLETINS 
 
 NUMBER II 
 
 SCALES FOR THE MEASUREMENT OF ENGLISH 
 COMPOSITIONS 
 
 PURPOSE AND METHODS OF THIS STUDY 
 WHAT THIS MONOGRAPH is 
 
 AN experiment with the Hillegas Scale l showed that the use of 
 such an objective measure did unify the grades given to composi- 
 tions by teachers. It was also found, however, that the Hillegas 
 Scale was not satisfactory to the teachers of Newton, owing to 
 what seemed to them to be inherent faults. These faults may be 
 stated briefly as follows: (i) the scale aims to measure too varied 
 a product; (2) the compositions in it are not typical of good 
 school work; (a) some are artificial, (b) others are " bookish," 
 really reproductions, 2 and (c) no conversation is contained in any 
 of them. Hence it was proposed to prepare another scale, or 
 rather a series of scales, in such a way as to obviate these apparent 
 defects. The following monograph is a report on the preparation 
 of such a series of scales. These scales were prepared through the 
 cooperation of the Joseph Lee Fellow for Research in Education, 
 the eighth grade teachers and the elementary school principals in 
 the public schools of Newton, assisted by the teachers of English 
 
 1 A Scale for the Measurement of Quality in English Composition by Young 
 People M. B. Hillegas. In Teachers College Record, September, 1912. 
 
 2 As an experiment we prepared a scale for measuring compositions in repro- 
 duction written by eighth grade pupils. After careful consideration of the matter 
 at a teachers' meeting, it was agreed that such a scale ought not to be published. 
 The reasons briefly stated are as follows: i. At best Reproduction is the least 
 desirable form of composition. 2. Reproduction as a form of theme writing is too 
 often characterized by a slavish repetition of words or of ideas. 3. The publication 
 of a standard scale is likely to emphasize this undesirable practice. 
 
 5 
 
6 HARVARD-NEWTON BULLETIN 
 
 in the high schools of Newton, and by teachers and principals in 
 Arlington, Massachusetts, and Boston. The teachers and prin- 
 cipals in Newton secured from their pupils the compositions used 
 in the experiment, and they graded and ranked each set of com- 
 positions. It was on the basis of their composite judgment that 
 the compositions were selected for each scale. The teachers of 
 English in the high schools of Newton assisted in denning the 
 merits and defects of the compositions in the scales. The 
 eighth grade teachers and elementary school principals in Arling- 
 ton, and a committee of similar teachers and principals in Boston 
 assisted in the experiment by testing the use of the description 
 scale after it had been finally prepared. Although the scales 
 were prepared primarily for use in the eighth grade in the Newton 
 Schools, they are published in order that whatever worth they 
 possess may be shared with teachers elsewhere. 
 
 THE PURPOSE or THE STUDY 
 
 The need of scales and similar standards as objective means of 
 measurement in education, is speedily gaining recognition as the 
 present insistent demand for educational efficiency increases. 
 The successful measurement of educational efficiency depends on 
 the establishment of standards of measurement in all depart- 
 ments of school work. The need of standards by which to meas- 
 ure the class-room work of pupils has been shown by every 
 investigator who has examined the present methods of rating or 
 grading school work. 1 Every such investigation has shown start- 
 ling variations in the grades given the same piece of work by 
 different teachers. Inasmuch as the piece of work graded is a 
 constant factor, the variation must be in the teachers who do the 
 grading. The variation may be due to the lack of a uniform or 
 common standard of judgment, or to the varying degrees of 
 
 1 For example, see Starch and Elliott, " Reliability of the Grading of High School 
 Work in English," School Review, Vol. 20, pp. 442-457. Also " Reliability of 
 Grading Work in Mathematics," ibid., Vol. 21, pp. 254-259. Also " Reliability of 
 Grading Work in History," ibid., Vol. 21, pp. 676-681. Also Jacoby, " Notes on 
 the Marking System in the Astronomical Course in Columbia University, 1909-10," 
 Science, Vol. 31, p. 819. 
 
HARVARD-NEWTON BULLETIN 7 
 
 accuracy with which teachers apply a common standard. In 
 view of the fact that the progress of the student throughout his 
 whole educational career is contingent on his having earned 
 acceptable grades from each successive teacher, it is clear that 
 the importance of accurate measurements of his intellectual 
 attainments cannot be over-emphasized. 
 
 It is to create an objective standard for the measurement of 
 one of these intellectual attainments, namely, the writing of 
 English, that the preparation of this series of scales is undertaken. 
 By the wise use of such standards it is hoped that teachers may 
 come to agree more uniformly in their judgments of the value of 
 an English composition written by an eighth grade pupil. It 
 may be reasonably doubted whether it will ever be possible for 
 educators to measure the results of school training as accurately 
 as laymen measure temperature with a thermometer, or a board 
 with a square, or a piece of carpeting with a yard stick. Never- 
 theless, objective standards should serve as a basis for the exercise 
 of the subjective judgment, and until objective standards of 
 measurement are established in the realm of educational practice 
 educational progress will be determined as heretofore, by mere 
 opinion, and hence cannot be based on scientifically determined 
 educational facts. 
 
 PRINCIPLES ON WHICH THE PREPARATION AND USE OF A 
 SCALE SHOULD BE BASED 
 
 From the attempted but only partially successful use of the 
 Hillegas Scale, it seemed that to be successfully used a scale for 
 measuring compositions must be formed and applied according to 
 the following general principles. 
 
 i. A scale should not aim to measure too complex a product. 
 To attempt to measure the several forms or types of English com- 
 position by one and the same scale is like trying to measure heat, 
 light and color by the same instrument. To measure success- 
 fully the common but different forms of English composition 
 written by any given group of pupils we need as many different 
 scales as there are different forms of compositions. This study 
 
8 HARVARD-NEWTON BULLETIN 
 
 includes the preparation of scales for measuring description, 
 exposition, argument, and narration. 
 
 2. A scale should be made up of compositions having the same 
 characteristics as those which the scale is intended to measure. 
 The qualities of a composition of a freshman in college are dif- 
 ferent from those of an eighth grade pupil although each student 
 may have written on the same subject, and although further each 
 composition may merit the same grade. Hence, each scale in 
 this study is made up of compositions actually written by eighth 
 grade pupils and is intended to be used in judging the quality of 
 compositions of eighth grade pupils. It is possible and indeed 
 probable, that these scales may prove almost equally serviceable 
 in the seventh grade and in the first year high school. 
 
 3. A scale should conform, as far as practicable, to the present 
 almost universal practice of rating school work either by the use 
 of letters or of numerical grades. There seems to be little, if any, 
 practical need of a scale for the eighth grade which shall measure 
 any work which is much below 40%. Hence, each of the follow- 
 ing scales contains six specimens or samples, of English composi- 
 tion graded A, B, C, D, E, and F, and representing in general l 
 the grades, 95%, 85%, 75%, 65%, 55%, and 45%, respectively. 
 
 4. The readers on whose combined judgments the selections of 
 the scale is based should be those persons most familiar with the 
 qualities of the compositions to be measured by the scale. This 
 is essential because the scale must find acceptance among those 
 who use it. Accordingly, the scales here presented were selected 
 on the basis of the combined judgments of eighth grade teachers 
 and such others as supervise, and hence are familiar with, eighth 
 grade work in English composition. A scale which has been ever 
 so scientifically prepared, but which does not carry conviction of 
 the possibility of its usefulness is worthless as a practical instru- 
 ment in the hands of the teacher. The piling up of numerous 
 
 1 " In general," because each composition in the scale is expected to form the 
 standard for measuring all compositions which fall within the limits of that ten 
 per cent of which the composition in the scale forms the center. For example, the 
 " A " grade composition is to be the standard for measuring all compositions which 
 vary in quality from 90% to 100%. 
 
HARVARD-NEWTON BULLETIN 9 
 
 judgments, of the contents of which nothing is definitely known, 
 does not appear to be a proper basis for selecting a scale. 
 
 5. The compositions of a scale must be analyzed as to merits 
 and defects, else there is not only no guarantee, but little prob- 
 ability, that the users of the scale will interpret the qualities of 
 those compositions any better than they now interpret the 
 qualities of compositions without the use of any objective scale. 
 Even though two individuals may have given a composition the 
 same grade, say 80%, there is no certainty that they judged the 
 composition on the same qualities. In fact, it is altogether prob- 
 able that they did not grade it on the same elements or qualities, 
 and the giving of the same grade to it does not, by any means, 
 indicate uniformity of judgment. Hence, merit must be defined, 
 or the scale remains to most would-be users merely a series of 
 compositions like any others. This definition we have tried to 
 make. The same teachers who marked the compositions have 
 defined those qualities of merit and defect which prompted their 
 ratings. The definition of the qualities of merit and defect have 
 been attached to each composition in each scale. By this means 
 each teacher who uses the scale will know what the readers con- 
 sidered the merits and defects of each composition. Without 
 such a definition of qualities, there is no certainty that the user 
 of the scale will see in each composition the same merits and 
 defects that the reader saw. Unless the user does recognize the 
 same merits and defects, there is little or no advantage in his 
 comparing a set of compositions to be graded with those of the 
 scale : he might as well compare them with one another. 
 
 6. Teachers must accept a scale as a standard, in the same 
 sense that they accept other standards. If one judges that a 
 piece of string is a foot long, and on measuring it with a tape 
 measure finds it is only ten inches, he accepts the correction of his 
 own subjective judgment as indicated by the standard tape meas- 
 ure. So it should be in measuring compositions. If a scale is to 
 become a real standard, it must be accepted as such and used in 
 the true sense of a standard. Each composition in the scales 
 which have been prepared has been selected on the basis of the 
 
io HARVARD-NEWTON BULLETIN 
 
 combined judgment of a group of well-qualified readers. The 
 assumption is that the combined judgment of a group is more 
 reliable than the single judgment of any member of the group. 
 When a composition is to be graded it is compared with the scale, 
 and if it is about like the 75% composition of the scale, that 
 should be the grade for the composition. Unless this is the case, 
 the scale cannot be called a standard of measurement. 
 
 To summarize: we have prepared a series of four scales with 
 which to measure respectively the four forms of composition- 
 writing of eighth grade pupils, namely, description, exposition, 
 argument, and narration. Each scale is made up of compositions 
 having the same characteristics as those which the scale is in- 
 tended to measure. The compositions in the scale were selected 
 on the basis of grades given by teachers and other school people 
 familiar with what eighth grade compositions ought to be. The 
 compositions in the scale range approximately by regular degrees 
 of quality from the very best which pupils of the eighth grade can 
 write to one about as poor as ought to be found in eighth grade 
 work. The compositions are defined as to merits and defects by 
 the same teachers who rated them, and who are to use them. 
 Although a scale made on this basis seems to promise much 
 beyond what has been thus far attempted, nevertheless, its real 
 worth will have to be determined in practical use in the school- 
 room. 
 
 SECURING THE COMPOSITIONS FOR THE EXPERIMENT 
 (a) Obtaining the Compositions from Pupils 
 
 The compositions used in this study were written by eighth 
 grade pupils in the schools of Newton, Massachusetts, as a part 
 of their regular school work. Several topics in description, 
 exposition, argument, and narration familiar to the pupils, were 
 suggested by the teachers or proposed by the pupils, and each 
 pupil chose his subject. Pupils were asked to write a com- 
 position " about a page in length, not over two pages." They 
 were allowed opportunity within school time, up to one hour, for 
 self-preparation for the writing, and another hour for self-correc- 
 
HARVARD-NEWTON BULLETIN n 
 
 tion of their productions, in order that each composition written 
 might be the best unaided writing of which each pupil was ca- 
 pable. By obtaining compositions written by pupils as regular 
 school exercises compositions were secured which are characteris- 
 tic of the work of eighth grade pupils in this school system. 
 
 In order to secure compositions representing the range of 
 quality characteristic of the work of eighth grade pupils, and at 
 the same time make as little unnecessary work as possible for the 
 teachers and the pupils, a selection of compositions was made in 
 the following manner. The eleven schools were grouped into 
 five groups having approximately the same number of eighth 
 grade pupils in each group. The first and fifth groups consisted 
 of one school each, the second, third and fourth of three schools 
 each. Each of these five groups secured from its pupils and 
 furnished for this study, specimens of a given type of composition: 
 for example, group one furnished narrations, group two, argu- 
 ments, and so forth. Each school furnished samples to the 
 number of 25% of its entire enrollment in the eighth grade. 
 Each class teacher was asked to select that 25% of the composi- 
 tions which represented progressively all degrees of ability, 
 including the best and the poorest. After the compositions had 
 thus been selected by each class teacher, they were numerically 
 graded by the eighth grade teachers l in the school and the prin- 
 cipal independently. 
 
 In order to increase the chances of getting compositions of suffi- 
 ciently superior quality to merit the highest grade, namely "A" 
 or 95%, each of the eleven schools in which there were eighth 
 grade pupils, sent in from one to three of its " best " compositions 
 in all forms of writing as judged by the eighth grade teacher or 
 by teachers and the principal. 
 
 This study has to do then with the " best " compositions 
 written by pupils in the eighth grade in all the schools of Newton, 
 as selected by the teachers and principals, and a selection of 
 
 1 In two cases the eighth grade teacher taught seventh grade: in one case, 
 where there was only one eighth grade teacher, the seventh grade teacher was 
 included in this preliminary grading. 
 
12 HARVARD-NEWTON BULLETIN 
 
 compositions made by the teachers from all the compositions 
 written by all the eighth grade pupils, in a school or groups of 
 schools, such selection including at least 25 % of all compositions 
 and also made with a view of securing compositions representing 
 all degrees of quality. 
 
 The number of " bests " sent in each of the series averaged one 
 for each school, and the number of others ranged from twenty- 
 two to twenty-six samples. Twenty-five samples in each series 
 seemed a sufficient number to be read and graded by teachers 
 as a basis for selecting a scale of six compositions representing 
 the grades A, B, C, D, E, and F. From the teachers' selections 
 of from thirty to forty compositions, it became necessary to 
 select the twenty-five to make up the set to be read. The basis 
 for this selection was the preliminary grades given each composi- 
 tion by the teachers and the principals, together with the judg- 
 ment of the director of the experiment. The director's judgment 
 was based on a careful reading of each composition, on a compar- 
 ison of the ratings given the compositions by the different readers, 
 and the general fitness of each composition for a place in the 
 scale. 
 
 Some compositions were obviously unfit for use in a scale : for 
 example, they were too long or too short: of course such were 
 rejected. On the above mentioned basis those compositions were 
 selected for final readings by teachers which would best serve the 
 purpose of the study, and which were typical of the grades, A, B, 
 etc., which were to be the basis of the scale. The twenty-five 
 compositions which finally made up the set from which a scale 
 was to be selected usually included about ten " bests " and 
 fifteen others, or three samples each of the representative grades 
 below 95 %. 
 
 (b) Preparing Compositions for the Readers 
 
 To prepare the compositions for readers, each composition was 
 reproduced in typewritten and mimeographed form exactly as it 
 had been written by the pupil. Extreme care was exercised to 
 reproduce each error of whatever kind, and to retain all the 
 
HARVARD-NEWTON BULLETIN 13 
 
 qualities or characteristics of the original composition. 1 At the 
 head of each composition as it was written by the pupil were, the 
 name, the age in years and months, and the school grade of the 
 pupil, together with the name of the school and the date. When 
 the compositions were mimeographed, each one was given a 
 number and all the above marks of identification were omitted. 
 By removing all marks of identification, each composition stood 
 before each reader on exactly the same basis, and the chances 
 were increased that it would be judged according to its inherent 
 and intrinsic worth. 
 
 (c) Instructions to Readers 
 
 After the twenty-five compositions of each type had been thus 
 prepared, one set was sent to each of the fourteen eighth grade 
 teachers, and to each of ten elementary school principals. 
 Instructions were given that each reader was to grade the 
 compositions wholly independently. 
 
 Here are the instructions for grading the compositions : 
 
 In rating it is suggested that the following procedure be uniformly 
 adopted : 
 
 (a) Arrange the themes in a series in the order of their merit. 
 
 (b) To the theme considered to be the best give the arbitrary rating of 
 
 95%- 
 
 (c) Rate each of the remaining themes with reference to this standard, 
 giving it a percentage value and entering the same on the sheet provided, 
 both under " No. of Theme ", and " Choice of Theme ". 
 
 It should be borne in mind that these samples are eventually to serve as 
 standards for eighth grade themes: it is important, therefore, in rating that 
 all other considerations be set aside and that an " F ", for example, be a 
 reasonable eighth grade " F ", etc. 
 
 Fixing the grade of the best composition at 95 % needs explana- 
 tion. It was the feeling of those who organized this study that 
 95 % rather than 100 % might fairly represent the highest achieve- 
 ment of any eighth grade pupil in writing a composition. Fur- 
 ther, it also seemed desirable that each reader should start from 
 the same numerical starting point in arranging the compositions. 
 
 1 Obviously, such qualities as neatness, legibility of handwriting, etc., were 
 eliminated. 
 
14 HARVARD-NEWTON BULLETIN 
 
 By asking each reader to give 95 % to the best composition, this 
 desirable end would be secured. No lower limit was fixed, but 
 the Newton readers knew that compositions worth much less than 
 about 40 % were not to be furnished by schools for the experiment. 
 This tended to fix the lower limits of the grading around 40%. 
 That it only tended to do so, and did not absolutely do so, is 
 indicated by the fact that compositions were graded as low as 
 20%. It was not the original intention to fix any other lower 
 limit than the usual zero. 
 The following is the form of rating sheet: 
 
 Ratings of Themes in Description 
 
 For convenience in handling the ratings, readers are asked to enter their 
 results in two columns: first, under the serial number of the theme give the 
 rating, and second, under the order of choice give both number and rating, 
 thus: (i) 85% and (i) 18-95%. 
 
 Number of Theme Choice of Theme 
 
 i 
 
 2 
 
 3 
 
 Etc. Etc. 
 
 THE DESCRIPTION SCALE 
 THE COMPOSITIONS USED 
 
 Following is the set of twenty-five compositions in description 
 used in this experiment: they were sent to the readers in this 
 form. 
 
 i 
 
 BELFAST, MAINE 
 
 The little town of Belfast is situated on the Penobscot Bay in 
 Maine. Wishing to see all we can of this place and its surround- 
 ings we will climb to the belfrey of the Methodist Church. Our 
 eyes fall first on the water which sparkles clear and blue on a 
 summer's day. The shore, where children love to play is not a 
 white sandy beach, but one of rocks and ledges. 
 
 Then we slowly turn our eyes to the town itself. They rest 
 upon the churches, the post office a square red brick building, the 
 
HARVARD-NEWTON BULLETIN 15 
 
 lumber mill and the shoe factory. Farther along we notice the 
 steamboat wharf with one of the Boston boats just landing; and 
 way over the other end of the town is a Sardine factory. Off in 
 the distance stretch the green fields and blue hills. This surely 
 is an ideal place to spend a summer vacation. 
 
 2 
 THE POLEPHEMUS CATERPILLAR 
 
 There are three different kinds of Polephemus caterpillars, the 
 one that I am going to describe is the largest. When he is at rest 
 he measures three inches long, and three quarters of an inch in 
 diameter. His color is green, and he has seven yellew bands up 
 and down the side. At a distance he looks like a leaf. There is 
 another line on the underside which runs from his head to his tail. 
 There are silver spots on these lines which sparkle in the sun. He 
 has the largest head of all the caterpillars it is about the size of a 
 persons little finger nail. He has two pairs of powerful jaws, with 
 which, if he is disturbed he will make a little click, like that of a 
 watch. Every time he does this he makes gestures with his ugly 
 head which makes him look quite fierce. His neck is something 
 like some kodaks he can reach way forward and then close up 
 again. 
 
 He has four pairs of prolegs which are very powerful. There 
 are three pairs of true or butterfly legs or the legs near his head. 
 His eyes are in two groups each group containing five eyes at the 
 bottom of his head they are very strong for an insect. With 
 them he can see seven or eight inches ahead of him. 
 
 Its feet have little have little hairs on them which enable him to 
 get a strong grip on a leaf or twig. All of this description I have 
 got my-self by studying them. 
 
 3 
 THE BOY'S CLUB 
 
 In the town of Auburndale part of Newton is where club is 
 situated to get there from the square you walk up Auburn Street 
 till you get to the third street on the right this is Woodbine Ter- 
 
16 HARVARD-NEWTON BULLETIN 
 
 race. Our club is in the second hous up in the garret it is a new 
 room and in comparitivly good condition as you enter you see a 
 book-case & a chair on the walls ther are rows of picturs in the 
 back of the room there is a window and a cot on the right side 
 there is a graphaphone and a box of records. 
 
 4 
 THE LIGHTHOUSE 
 
 Twilight was falling as I strolled along the broad, sandy beach, 
 now and then stumbling over a projecting rock. 
 
 Nothing was to be heard but the dull booming of the breakers 
 on the over-hanging rocks and low beach, and the sad call of the 
 gulls in their flight home-ward. 
 
 Something in the atmosphere told me a storm was brewing. 
 A fog was gathering. 
 
 The rocks grew more rough and jagged as I advanced, and a 
 tall, gray lighthouse loomed up in front of me, high above the 
 rest of the scenery. Its great revolving light cast a warning ray 
 across the angry water. 
 
 In the rear of this tall lighthouse rose a forest of pines, such as 
 are found all along the Maine coast. Down the beach a little 
 way was moored a large row-boat. On the right lay a little cove, 
 like a haven to many a weary sailor. 
 
 As the fog grew heavier the lighthouse looked like a great 
 guardian angel warning the sailors that dangerous rocks would 
 wreck their vessels if they came too near. 
 
 5 
 A LIGHTHOUSE 
 
 Right near my house there is a very high sort of a cliff. On the 
 top of the cliff sets this very large lighthouse, it is mad of bricks 
 which are painted white. It is a very high and wide one com- 
 pared with most of the lighthouse. A little way down the side 
 cliff they have made some kind of a construction to blow the fog- 
 horn. On the top of this little building there is a very large bell 
 which they use for different things. In back of the light there is 
 
HARVARD-NEWTON BULLETIN 17 
 
 quite a good sized house which is a part of the lighthouse. In 
 this house the keeper of the light lives and his family. 
 
 6 
 AN EXCITING FIRE 
 
 " Fire! fire! " was the cry. Everybody was hustling and 
 bustling about a peaceful country town. Fires were a very 
 uncommon thing in that vicinity and the people were greatly 
 aroused. 
 
 Then in the distance the faint sounds of the engines were heard. 
 Louder and louder it grew until at last the engines were seen. 
 
 The people ran and the boys jumped on their bycicles for they 
 wanted to see the fire before it was out. The crowd were gather- 
 ing about a hay and grain barn. 
 
 Then as the fire engines sped around the corner the great hose 
 was hitched to the hydrant. The firemen worked hard and in a 
 few minutes the hose was pouring forth all its contents on the 
 blazing barn. 
 
 7 
 A CHRISTMAS MORNING 
 
 It was nine o'clock Christmas morning when a family of three 
 children and their parents were entering the room where the 
 Christmas tree stood. When they entered there were many 
 shouts and excited faces. The tree was a tall one which reached 
 to the ceiling it was covered with many glittering ornaments. 
 Underneath the tree were all of the presents which were to be 
 opened one at a time by the merry group. There were many 
 games played. 
 
 After this they all went into the dining room, where the good 
 smell of turkey and plum pudding existed. 
 
 8 
 THE WOODS IN WINTER 
 
 The tiny pool, shaded in summer by thriving young birches, 
 was now a crystal mirror. The rays of the morning sun gleamed 
 
i8 HARVARD-NEWTON BULLETIN 
 
 through the bushes, which sparkled like millions of diamonds. 
 Their crust of ice had been formed in the night, and at dawn, the 
 sun rose over a glistening world, for every twig and limb shone in 
 a coat of ice or snow. 
 
 The pines, in the wood, were robed in white, and dazzled in 
 their glittering garments. All nature gleamed in transparent 
 whiteness, for the earth was one sparkling gem. 
 
 9 
 THE APPROACHING TRAIN 
 
 It was quiet dark when the people were around the depot 
 waiting for the next train to come. It was a cold and windy 
 night for the people, the hackmen ran around and did everything 
 to keep warm. 
 
 They were about thirty-five men, and about fifteen women, 
 and a few boys. These people were all waiting for a train to go 
 home. It was a cold night for the boys and they played tag to 
 get warm. 
 
 The men and boys that came to the depot to get the papers and 
 bundles had to hurry home, because of the cold weather, just 
 about five minutes later, I heard the whistle. The bell in the 
 depot began to ring, and I saw the saw the train just taking the 
 curve, when it reached to the depot, all the people were glad to get 
 in and get warm, and the baggage-master gave the signal to start, 
 and the train started off. 
 
 10 
 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 fore- 
 telling the coming of a storm. The houses across the way were 
 dismal shadows, flat, cold, heartless. A piercing chill pene- 
 trated 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 
 
HARVARD-NEWTON BULLETIN 19 
 
 cold everything was cold, cold, cold. An awful lonliness 
 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, another, and then a 
 continuous blowing. The howling was 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 
 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 extinguished, and the 
 roaring called to mind the last day at 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 indoors, to the poor shelter afforded 
 by the groaning houses. 
 
 For about an hour the storm continued thus, then by degrees 
 the wind lessened, though the rain still fell, and the ocean thun- 
 dered. 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 town and the great piles of wreckage 
 on the shore. 
 
 ii 
 MY FAVORITE PET 
 
 My favorite pet is a horse which I am very familiar with. He 
 is a very large horse with a large chest. It is very old and his 
 color is gray and white but mostly white. Its teeth are very 
 sharp and big. His main is cut short and rounds at the top. His 
 neck is long and very broud. He has little black spots around 
 his neck from wearing a big collar. When you fool with him he 
 would try to bite you. 
 
20 HARVARD-NEWTON BULLETIN 
 
 His tail is long and white which some parts are gray. Its hoof 
 is very hard and if you hit it with a hammar you could not hert 
 him. Its legs are of a dapple gray. Around which the shoe goes. 
 Its mouth it is very black. By his hoof there is a black spot 
 where he was cut by a barb wire. He has a very small hoof and 
 his heir is very long at the hoof. 
 
 12 
 A COUNTRY VILLAGE 
 
 There is a very old village called East Foxboro. Just as you 
 get of the train the first thing you see is a quiet little village with 
 railway tracks running through the centre, on one side is a large 
 depot as large a city depot. Just beside that is a small store and 
 post-office all in one. Joined on to that is a barn. On the other 
 side is a freight-house with running to it. In back of that is a 
 dry-good, newspaper, candy, tonic, and postcard store. Not far 
 from that is a pump and wooden box in which the water falls for 
 horses. Nearby is a shed in which gasoline, oil and kerosene is 
 sold. Just across the well worn road is a signal station which is 
 largely used but very old. 
 
 13 
 
 GRANDMOTHER 
 
 In front of the open fire place 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 carefully tied. Her face is full 
 of wrinkles and her keen blue 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 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 into her work-basket. 
 
HARVARD-NEWTON BULLETIN 21 
 
 Then she trots upstairs to bed and oh, how lonesome it is when 
 our dear Granny is gone from the room. 
 
 14 
 THE DISCRIPTION OF A PLACE IN MAINE 
 
 As we stand in the door of the log cabin and directly across a 
 lake which is about fifty yards away from the front of the log 
 cabin we see some of the woods of Maine, and as we look a little 
 to the left of the lake Squaw mountain may be seen. On clear 
 days the top of the mountain can be seen and sometimes the Old 
 Squaw's face can be seen (it is of stone). On rainy days only the 
 outline of the lake can be seen, and sometimes it is so very foggy 
 that the lake can hardly be seen from the door of the of the 
 cabin. Sometimes a deer can be seen very early in the morning. 
 
 THE VIEW FROM MX. WASHINGTON 
 
 The clouds parted, and, far below us, the country was bathed ) 
 in sunshine. Everything was very small. The Mt. Washington 
 hotel looked like a cottage, though it was the largest building for 
 miles around. 
 
 Here and there, small farm houses dotted the landscape. 
 Occasionally a river, or stream, wound its way, like a blue thread, 
 through the green meadows. In the meadows nearest us, small 
 black specks could be seen. With the aid of field glasses, the 
 specks proved to be horses and cows. 
 
 The roads looked like yellow threads, along which, black 
 specks moved. Looking from the other side of the mountain, the 
 railroad which came up the mountain, could be followed. 
 
 Other mountain tops surrounded us on all sides. There was 
 one valley, that was so completely surrounded by mountains, 
 that there seemed no way to leave it. 
 
 The whole country was clothed in green, while the sky was 
 blue, with fleecy clouds floating lazily across it. Clouds of 
 smoke, showed here and there against the green hills. Sud- 
 denly the clouds came to-gether, and, we saw no more. 
 
22 HARVARD-NEWTON BULLETIN 
 
 16 
 A PICTURE 
 
 On the wall in the front of the schoolroom hangs a large picture, 
 with a frame of dark wood. The name, I am not sure of. 
 
 In the background is a dense forest with an opening at the left 
 hand corner, where, far away in the distance, can be seen the 
 beautiful tinted sky. The ground is covered with snow, and in 
 the foreground a few trees and boulders are scattered here and 
 there. 
 
 The centre of attraction is the number of Puritan people walk- 
 ing in the opening. Some men carry guns in defense from the 
 Indians. Many small children are by their mother's side, and 
 most all carry a Bible. 
 
 The men are dressed in brown doublets, wearing short, knee 
 length trousers, with leggings and brown leather boots, stiff 
 white collars and cuffs, gray hats, with high crowns and wide 
 brims, carrying guns across their shoulders. 
 
 The women wears capes with attached hoods, some green, 
 white and yellow. Some wear the black bonnets with a white 
 band on the edge. In the right hand corner more men are coming 
 along. 
 
 From this description I should name this characteristic picture, 
 " Puritans on their way to church." 
 
 17 
 FLANDERS LAKE 
 
 There! see ? the blue lake just at the end of the path, and 
 there 's the big ledge on which there are so many comfortable 
 nicks and crannies to sit in. Way across this lake is Sckoodic, 
 grand old mountain, with the little camps scattered about at it's 
 feet. 
 
 Come, let 's sit down and rest. Here and there rocky points 
 jut out into the lake terminating in massive boulders. Between 
 these points are beautiful stretches of coarse sand that glisten 
 like so many jewels and particles of gold and silver in the sun. 
 
HARVARD-NEWTON BULLETIN 23 
 
 Groves of young birches growing close to the water's edge are 
 dotted here and there with small pines and spruce. Back of these 
 birches is a veritable forest of pines, spruce, balsam, and other 
 varieties of evergreen which rend the air fragrant with their 
 soothing odor. 
 
 But see, it is getting dark and we had better start our walk 
 home through these fragrant woods for even the busy king-fisher 
 has stopped his rattle. 
 
 18 
 
 AN OLD-FASHIONED FARMHOUSE 
 
 On a dusty country road in Concord, which is not far from here, 
 there stands an old-fashioned farmhouse. The roof slopes down 
 to the porch over which a honeysuckle vine grows. The clap- 
 boards are stained and weatherbeaten. The windows are small 
 and have small panes in them. On them are queer, old- 
 fashioned shutters. 
 
 At one side of the house there is an old-fashioned barn which 
 appears to be large and airy but is much in need of repair. Some 
 of the windows are broken and are filled with hay, and the doors 
 after many years of hard service are partly off their hinges. The 
 once gilded weather-cock has changed its gilt to rust. 
 
 When one thinks of the house and barn to-gether it reminds 
 them of the comforts of the old days. 
 
 A SILENT STRANGER 
 
 All is still on the lake ? Oh no! A fish leaps out of the dark 
 water his scales glissning in the moonlight. All is still in the 
 forest behind ? No! A rabbit jumps out of the ferns on the 
 bank, stares a moment, then leaps into the bushes. 
 
 Ah! What is that so silent in the shade over there ? A tall 
 youth clad in buckskin with fringe on his shirt and leggings, a 
 buff alow horn on one side and a buckskin pouch on the other held 
 on by a leatheron thongs over his broad shoulders, broader than 
 eny I had ever seen before, a skilfully decorated sheath in which 
 
24 HARVARD-NEWTON BULLETIN 
 
 was a sharp knife fastened to a buckskin band around his waist. 
 Held carelessly in his arms was a long-barroled rifle. He was 
 sturdy, strong and fearless, deeply tanned, with black eyes so 
 sharp that they could pearce the darkness of the forest. 
 
 I looked away for a moment. When my gaze was again 
 directed on the spot where I had seen the silent stranger he dis- 
 sipeared. Heavy as he was not a twig snaped or a leaf rustle 
 beneath his feet. 
 
 About fifteen munites after when I was looking along the shore 
 I saw a canoe glide silently out on the lake headed for the other 
 shore. In that canoe was the silent stranger who had so mysteri- 
 ously disiapeared. Not a sound came from his paddle as the 
 canoe sped swiftly along. 
 
 20 
 AN OLD SCHOOL HOUSE 
 
 It was a low studied old School-house, which stood on the 
 country road between Wissing and Eustace, New York. Win- 
 dowless and partially roofless it stood, as though it had nary a 
 friend with the exception of the cold winter air, as it leaped 
 boldly through the bare cracks in the side of the structure. The 
 door, which was nearly hidden from view, was swinging solemnly 
 back and forth on two rusty hinges. Really it was only one, as 
 the other was merely a piece of worn-out iron, which dangled 
 loosely in the draft of the swaying door. The old wooden desks 
 could no more be called varnished, as the dust, and the numerous 
 inks tains on them, hid all signs of it from observation. Yet, by 
 the various initials which were carved on the wood-work, it was 
 very evident that at one time it was not forsaken. The old 
 clock on the wall was motionless, and the hands which had once 
 explained the time, no longer looked like hands, but like little 
 ink spots across the forlorn face. Then there were the black- 
 boards, covered with pictures of every description, with the 
 scribbled signatures of teachers, written rudely under them. 
 The old text books with no covers lay stranded on the wornout 
 floor, and the stove, which once had been used for warmth, stood 
 
HARVARD-NEWTON BULLETIN 25 
 
 idely on top of the old rug which had, long years ago, served for a 
 carpet. 
 
 Such was the appearance of the first country schoolhouse I 
 ever was in. 
 
 21 
 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 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. 
 
 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 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. 
 
 22 
 THE LAKE AT SUNRISE 
 
 In the Mountains of Pennsylvania there is a little 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 the sun peeping up over the 
 top of the mountains and shinning on the water. Than 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 and 
 coves backed up by woods from which comes the chirp of the 
 
26 HARVARD-NEWTON BULLETIN 
 
 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 with the 
 roof sparkling with. On the roof of the house a hawk is sitting 
 adding his clear whistles 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 in flocks, and 
 far away from the hills comes the sound of the of Italians singing. 
 
 23 
 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 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 accident. In side there 
 is ah enormous light which flashes every two minutes and some- 
 times 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. 
 
 24 
 
 A SMASH-UP 
 
 It was early one morning last spring, when an auto belonging 
 to Harry Atwood coming from New York, run into a car at 
 Auburn St. The auto was going at a good rate when it struck 
 the car. When the auto struck the car it made a hole in the side 
 of the car. 
 
 His auto was a red one and on each side it had his name in 
 black letters Harry Atwood. The lights were broken and the 
 
HARVARD-NEWTON BULLETIN 27 
 
 mud guard were too. The front of the auto was jamed. It was 
 a auto that would seat four people. With a cover to keep the 
 rain out. There was marks where they had pushed the auto off 
 the track. That afternoon it was carried away. 
 
 25 
 A SCENE ON THE PRAIRIE 
 
 Along a large plain in the west w^th 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 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 darkness 
 and the trappers turned into camp leaving one a the trappers on 
 gaurd. 
 
28 HARVARD-NEWTON BULLETIN 
 
 THE TABULATION OF RESULTS OF READINGS 
 (a) THE GRADE BASIS 
 
 When the readers made their reports showing the grades which 
 they had given each composition these grades were summarized 
 in the following table. 
 
 Table I shows the grade that each of the twenty-four readers 
 gave each of the twenty-five compositions in description. 
 
 TABLE I 
 
 Number of the Composition 
 / 23 4 s ^ 7 8 g 10 ii 12 13 14 15 16 17 18 19 20 21 22 23 24 25 
 
 75 80 40 90 50 77 65 83 48 95 50 55 86 45 88 70 76 73 78 85 76 47 45 30 20 
 
 72 75 42 75 60 63 53 90 55 95 47 62 82 58 92 85 75 70 68 80 75 65 57 52 45 
 83 64 50 87 52 65 '55 79 60 95 40 45 85 78 91 75 89 72 71 92 54 70 59 44 48 
 
 82 90 30 90 60 80 70 92 45 95 45 60 88 40 92 92 90 80 82 78 75 75 45 42 40 
 90 88 50 91 73 80 73 95 7o 95 55 75 85 70 93 80 85 90 75 88 75 65 58 60 45 
 
 75 85 45 85 55 75 60 70 75 90 65 65 95 45 85 85 85 80 90 80 80 65 65 45 55 
 
 83 78 40 63 60 75 75 80 59 95 50 60 75 52 90 80 82 79 78 85 78 73 58 61 52 
 
 73 80 45 85 60 60 55 90 60 95 50 60 85 50 90 80 70 75 70 80 80 60 65 50 55 
 83 85 40 70 50 55 65 60 50 95 40 60 75 40 90 70 63 73 92 87 80 77 60 40 40 
 85 80 40 75 65 50 60 70 50 95 40 60 75 55 90 70 70 75 90 90 78 75 60 45 40 
 83 84 50 90 68 80 70 93 65 95 58 73 87 65 83 77 81 87 78 85 80 75 65 60 50 
 80 70 45 85 65 65 58 88 45 95 50 6 90 60 93 68 78 80 75 85 70 53 45 55 45 
 83 82 40 94 78 88 80 84 68 95 50 75 90 65 93 91 92 87 81 85 89 86 74 73 45 
 95 88 50 85 70 78 70 95 60 95 50 65 80 75 90 80 88 90 85 88 80 70 55 58 45 
 92 85 45 86 58 76 74 92 60 95 45 70 85 58 86 69 83 74 78 70 73 7i 45 55 45 
 90 78 40 88 58 50 72 52 40 95 46 42 71 40 92 75 72 60 80 73 74 65 45 42 42 
 90 82 40 83 58 64 60 78 63 95 55 45 93 65 94 70 80 75 84 85 85 68 50 50 40 
 
 76 74 43 90 60 70 66 86 58 95 45 53 79 63 93 91 87 84 80 72 82 50 56 46 40 
 92 89 45 93 70 75 47 93 60 95 57 67 92 48 83 86 88 78 85 85 78 71 65 52 62 
 68 70 40 75 58 62 65 70 52 95 48 55 85 50 88 72 72 74 90 80 66 65 60 45 42 
 
 82 80 45 79 75 72 40 93 65 91 60 68 89 50 95 85 80 88 60 90 80 77 55 40 48 
 
 74 72 40 86 60 68 60 80 52 95 45 60 75 60 80 73 85 75 78 84 73 63 60 50 50 
 78 82 39 84 55 63 62 93 45 95 49 6l 82 55 89 80 93 73 80 78 71 62 40 38 44 
 
 83 79 40 93 53 69 70 70 42 95 49 42 80 45 90 73 80 66 83 79 75 So 42 42 40 
 
 This table shows that reader No. i gave composition No. i, 75 %: 
 composition No. 2, 80%: composition No. 3, 40%, and so on. 
 By following the vertical column of figures under any given com- 
 position (e. g. No. i) one can see the variety of grades given that 
 
 1 For obvious reasons, each reader was given a number. This number the 
 reader retained throughout this investigation. A reader may find out what his 
 number was by addressing the author of this monograph. 
 
HARVARD-NEWTON BULLETIN 29 
 
 composition. To facilitate such an observation, the following 
 table has been prepared. 
 
 Table II shows the grades given each composition in descrip- 
 tion arranged in a descending order without designating the 
 readers who gave them. 
 
 TABLE n 
 
 Number of the Composition 
 
 i a 34 5 6 7 8 Q 10 n 12 13 14 15 16 17 18 IQ 20 21 22 23 24 2$ 
 
 95 oo 50 94 78 88 80 95 75 95 65 75 95 ?6 95 92 93 9O 92 92 89 86 74 73 62 
 
 92 89 50 93 75 80 75 95 7o 95 60 75 93 75 94 9* 92 90 90 90 85 77 65 61 55 
 
 92 88 50 93 73 80 74 93 68 95 58 73 92 70 93 91 90 88 90 90 82 77 65 60 55 
 
 9o 88 50 91 70 80 73 93 65 95 57 70 90 65 93 86 89 87 90 88 80 75 65 60 52 
 
 90 85 45 90 70 78 72 93 65 95 55 68 90 65 93 85 88 87 85 88 80 75 65 58 50 
 
 90 85 45 90 68 77 70 93 63 95 55 67 89 65 93 85 88 84 85 87 80 75 60 55 SO 
 
 85 85 45 90 65 76 70 92 60 95 50 65 88 63 02 85 87 80 84 85 80 73 60 55 48 
 
 83 84 45 90 60 75 70 92 60 95 50 65 87 60 92 80 85 80 83 85 80 71 60 52 48 
 
 83 82 45 88 60 75 70 90 60 95 50 62 86 60 92 80 85 80 82 85 80 71 60 52 45 
 
 83 82 45 87 60 75 66 90 60 95 50 61 85 58 91 80 85 79 81 85 78 70 59 50 45 
 
 83 82 43 86 60 72 65 88 60 95 50 60 85 58 90 80 83 78 80 85 78 70 58 50 45 
 
 83 80 42 86 60 70 65 86 59 95 50 60 85 55 90 80 82 75 80 85 78 68 58 50 45 
 
 83 80 40 85 60 69 65 84 58 95 49 60 85 55 90 77 81 75 80 85 76 65 57 46 45 
 
 82 80 40 85 60 68 62 83 55 95 49 60 85 52 90 75 80 75 78 84 75 65 56 45 45 
 
 82 80 40 85 58 68 60 80 52 95 48 60 82 50 90 75 80 75 78 80 75 65 55 45 44 
 
 80 79 40 85 58 65 60 80 52 95 47 60 82 50 90 73 80 74 ?8 80 75 65 55 45 42 
 
 78 78 40 84 58 65 60 79 50 95 46 60 80 50 89 73 78 74 ?8 80 75 65 50 44 42 
 
 76 78 40 83 58 64 60 78 50 95 45 55 80 48 88 72 76 73 78 80 74 63 45 42 40 
 
 75 75 40 79 55 63 58 70 48 95 45 55 79 45 88 70 75 73 75 79 73 62 45 42 40 
 
 75 74 40 75 55 62 55 70 45 95 45 53 75 45 86 70 72 73 75 ?8 73 60 45 42 40 
 
 74 72 40 75 53 60 55 70 45 95 45 45 75 45 85 70 72 72 71 ?8 71 53 45 40 40 
 
 73 70 40 75 52 55 53 70 45 95 40 45 75 40 83 70 70 70 70 73 70 50 45 40 40 
 
 72 70 39 70 50 50 47 60 42 91 40 42 75 40 83 69 70 66 68 72 66 50 42 38 40 
 
 68 64 30 63 50 50 40 52 40 90 40 42 71 40 80 68 63 60 60 70 54 47 40 30 20 
 
 This table shows how each composition was treated by all 
 twenty-four of the readers. Under the number of the composi- 
 tion at the head of each column is given the various grades which 
 that composition received. For example, composition No. i 
 was graded 95% by one reader: 92% by two readers: 90% by 
 three readers, and so on. The highest grade given composition 
 No. i was 95%, and the lowest 68%. 
 
 Table III gives (a) the highest grade, (b) the lowest grade, 
 (c) the maximum variation, (d) the mean or average grade, and 
 (e) the median grade, for each of the twenty-five compositions in 
 description. 
 
HARVARD-NEWTON BULLETIN 
 
 It will be recalled that each reader was instructed (a) to 
 select the best composition and arbitrarily give it a grade of 95 %, 
 and (b) to arrange the remaining compositions in the set in a 
 descending order of merit. The purpose of the direction was to 
 make each reader start from the same numerical starting point. 
 In view of these instructions, and also in view of the fact that by 
 grading each composition the reader also necessarily had to ar- 
 
 TABLE III 
 
 Composition 
 Number 
 
 Highest 
 Grade 
 
 Lowest 
 Grade 
 
 Maximum 
 Variation 
 
 Mean or Average 
 Grade 
 
 Median 1 
 Grade 
 
 I 
 
 95 
 
 68 
 
 27 
 
 91.9 
 
 83.0 
 
 2 
 
 QO 
 
 64 
 
 26 
 
 8o.O 
 
 8o.O 
 
 3 
 
 50 
 
 30 
 
 2O 
 
 42-7 
 
 41.0 
 
 4 
 
 94 
 
 63 
 
 31 
 
 84.3 
 
 85.5 
 
 5 
 
 78 
 
 SO 
 
 28 
 
 6i.x 
 
 6o.O 
 
 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-o 
 
 ii 
 
 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 
 
 IS 
 
 89.6 
 
 90.0 
 
 16 
 
 92 
 
 68 
 
 24 
 
 78.2 
 
 78.5 
 
 17 
 
 93 
 
 63 
 
 30 
 
 81.0 
 
 8i. S 
 
 18 
 
 90 
 
 60 
 
 30 
 
 79-9 
 
 75-o 
 
 iQ 
 
 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-o 
 
 1 " Median grade " is the grade in the series of grades above which and below 
 which there is an equal number of grades. 
 
 For the definition of statistical terms used, and the methods of computation, see 
 Thorndike, Mental and Social Measurements, The Science Press, New York, 
 Whipple, Manual of Mental and Physical Tests, Warwick & York, Inc., Baltimore. 
 
HARVARD-NEWTON BULLETIN 31 
 
 range them in an order of merit, it is possible, in considering the 
 results, to use not only the grade basis but also the rank basis for 
 selection. From the grades given in Table II, we have computed 
 the maximum variation, and the mean or average grade and the 
 median grade as shown in Table III. As a further basis for 
 selecting the compositions for the scale, and as a check on the 
 above items, we have also computed the same items on the basis 
 of the ranking of each composition, i. e., on the basis of the 
 relative position of each composition among the twenty-five 
 specimens in the set. 
 
 (6) THE RANK BASIS 
 
 Each reader was directed to select the best composition and 
 arbitrarily to give it a grade of 95%. Since each of the twenty- 
 four readers did so, it is fair to assume that they gave equally care- 
 ful attention to the second direction, to arrange the other com- 
 positions in the order of their relative merit from 95% down. 
 There is no reason to believe that readers were either more or less 
 careful about the intrinsic or inherent worth of the compositions 
 as represented by the grades given them than they were about the 
 relative merits of compositions as indicated by their relative 
 ranks. The result of the ranking of the twenty-five compositions 
 by the twenty-four readers is given in the following table. 
 
 Table IV shows the ranking of each composition in description 
 by each reader. 
 
 This table shows that reader No. i selected composition No. to 
 as first in rank, composition No. 4 as second, composition No. 15 
 as third, and so on. Reading the table from top to bottom, one 
 finds in each column the different compositions that have been 
 assigned any given rank. For example, composition No. 10, is 
 first choice of twenty-one out of twenty-four readers. Reader 
 No. 6, selected composition No. 13 as first in rank, and reader 
 No. 14 selected composition No. 8 as first, and reader No. 21 
 selected composition No. 15 as first. 
 
 In order to show still more clearly the significance of the 
 material in this table, each composition has been dealt with by 
 
32 HARVARD-NEWTON BULLETIN 
 
 8-1 
 
 6 8 
 
 3 
 
 3 
 
 4 
 
 5 
 
 6 
 
 7 
 
 8 
 
 9 
 
 TABLE IV 
 
 Choice in Rank 
 
 IO II 12 13 14 
 
 15 
 
 
 17 
 
 
 19 
 
 20 
 
 21 
 
 22 
 
 23 
 
 24 
 
 25 
 
 z 
 
 10 
 
 4 
 
 IS 
 
 13 
 
 20 
 
 8 
 
 2 
 
 19 
 
 6 
 
 21 
 
 17 
 
 i 
 
 18 
 
 16 
 
 7 
 
 12 
 
 S 
 
 II 
 
 9 
 
 22 
 
 14 
 
 23 
 
 3 
 
 24 
 
 25 
 
 2 
 
 10 
 
 IS 
 
 8 
 
 16 
 
 13 
 
 20 
 
 2 
 
 4 
 
 i? 
 
 21 
 
 I 
 
 18 
 
 19 
 
 22 
 
 6 
 
 12 
 
 S 
 
 14 
 
 23 
 
 9 
 
 7 
 
 24 
 
 ii 
 
 25 
 
 3 
 
 3 
 
 10 
 
 20 
 
 IS 
 
 17 
 
 4 
 
 13 
 
 Z 
 
 8 
 
 14 
 
 16 
 
 18 
 
 19 
 
 22 
 
 6 
 
 2 
 
 9 
 
 23 
 
 7 
 
 21 
 
 5 
 
 3 
 
 25 
 
 12 
 
 24 
 
 ii 
 
 4 
 
 10 
 
 IS 
 
 16 
 
 8 
 
 17 
 
 2 
 
 4 
 
 13 
 
 19 
 
 Z 
 
 6 
 
 18 
 
 2O 
 
 21 
 
 22 
 
 7 
 
 S 
 
 12 
 
 II 
 
 23 
 
 9 
 
 24 
 
 25 
 
 14 
 
 3 
 
 S 
 
 10 
 
 8 
 
 15 
 
 4 
 
 18 
 
 I 
 
 2 
 
 20 
 
 17 
 
 13 
 
 6 
 
 16 
 
 19 
 
 12 
 
 21 
 
 S 
 
 7 
 
 9 
 
 14 
 
 22 
 
 24 
 
 23 
 
 II 
 
 3 
 
 25 
 
 6 
 
 13 
 
 19 
 
 10 
 
 4 
 
 IS 
 
 17 
 
 16 
 
 2 
 
 18 
 
 20 
 
 21 
 
 z 
 
 6 
 
 9 
 
 8 
 
 ii 
 
 12 
 
 22 
 
 23 
 
 7 
 
 5 
 
 25 
 
 3 
 
 14 
 
 24 
 
 7 
 
 IO 
 
 IS 
 
 20 
 
 i 
 
 17 
 
 16 
 
 8 
 
 18 
 
 21 
 
 2 
 
 19 
 
 13 
 
 6 
 
 7 
 
 22 
 
 4 
 
 24 
 
 12 
 
 5 
 
 9 
 
 23 
 
 14 
 
 25 
 
 II 
 
 3 
 
 8 
 
 IO 
 
 IS 
 
 8 
 
 4 
 
 13 
 
 20 
 
 2 
 
 16 
 
 21 
 
 18 
 
 I 
 
 19 
 
 17 
 
 23 
 
 22 
 
 12 
 
 6 
 
 9 
 
 5 
 
 25 
 
 17 
 
 II 
 
 24 
 
 14 
 
 3 
 
 9 
 
 IO 
 
 19 
 
 IS 
 
 20 
 
 2 
 
 I 
 
 21 
 
 22 
 
 13 
 
 18 
 
 4 
 
 16 
 
 7 
 
 17 
 
 8 
 
 23 
 
 12 
 
 6 
 
 5 
 
 9 
 
 25 
 
 24 
 
 14 
 
 II 
 
 3 
 
 10 
 
 10 
 
 20 
 
 19 
 
 IS 
 
 Z 
 
 2 
 
 21 
 
 22 
 
 18 
 
 13 
 
 4 
 
 17 
 
 16 
 
 8 
 
 s 
 
 7 
 
 23 
 
 12 
 
 14 
 
 6 
 
 9 
 
 24 
 
 3 
 
 II 
 
 25 
 
 iz 
 
 10 
 
 8 
 
 4 
 
 13 
 
 18 
 
 20 
 
 2 
 
 Z 
 
 IS 
 
 17 
 
 6 
 
 21 
 
 19 
 
 16 
 
 22 
 
 12 
 
 7 
 
 5 
 
 14 
 
 23 
 
 9 
 
 24 
 
 25 
 
 II 
 
 3 
 
 12 
 
 10 
 
 15 
 
 13 
 
 8 
 
 20 
 
 4 
 
 Z 
 
 18 
 
 17 
 
 19 
 
 2 
 
 21 
 
 16 
 
 6 
 
 S 
 
 14 
 
 12 
 
 7 
 
 24 
 
 22 
 
 ii 
 
 9 
 
 23 
 
 25 
 
 3 
 
 13 
 
 IO 
 
 4 
 
 15 
 
 17 
 
 16 
 
 13 
 
 21 
 
 6 
 
 18 
 
 22 
 
 20 
 
 8 
 
 I 
 
 2 
 
 19 
 
 7 
 
 5 
 
 12 
 
 23 
 
 24 
 
 9 
 
 14 
 
 Ii 
 
 25 
 
 3 
 
 14 
 
 8 
 
 i 
 
 10 
 
 IS 
 
 18 
 
 2 
 
 17 
 
 20 
 
 4 
 
 19 
 
 13 
 
 16 
 
 21 
 
 6 
 
 14 
 
 S 
 
 7 
 
 22 
 
 12 
 
 9 
 
 24 
 
 23 
 
 3 
 
 II 
 
 25 
 
 IS 
 
 10 
 
 I 
 
 8 
 
 4 
 
 IS 
 
 2 
 
 13 
 
 17 
 
 19 
 
 6 
 
 7 
 
 18 
 
 21 
 
 22 
 
 12 
 
 20 
 
 16 
 
 9 
 
 14 
 
 5 
 
 24 
 
 II 
 
 23 
 
 25 
 
 3 
 
 16 
 
 10 
 
 IS 
 
 z 
 
 4 
 
 19 
 
 2 
 
 16 
 
 21 
 
 20 
 
 7 
 
 17 
 
 13 
 
 22 
 
 18 
 
 5 
 
 8 
 
 6 
 
 II 
 
 23 
 
 24 
 
 12 
 
 25 
 
 9 
 
 14 
 
 3 
 
 17 
 
 10 
 
 IS 
 
 13 
 
 i 
 
 20 
 
 21 
 
 19 
 
 4 
 
 2 
 
 17 
 
 8 
 
 18 
 
 16 
 
 22 
 
 U 
 
 6 
 
 9 
 
 7 
 
 S 
 
 ii 
 
 24 
 
 23 
 
 12 
 
 25 
 
 3 
 
 18 
 
 IO 
 
 IS 
 
 16 
 
 4 
 
 17 
 
 8 
 
 18 
 
 21 
 
 19 
 
 13 
 
 Z 
 
 2 
 
 20 
 
 6 
 
 7 
 
 14 
 
 S 
 
 9 
 
 23 
 
 12 
 
 22 
 
 24 
 
 II 
 
 3 
 
 25 
 
 19 
 
 IO 
 
 4 
 
 z 
 
 8 
 
 13 
 
 2 
 
 17 
 
 16 
 
 19 
 
 IS 
 
 20 
 
 18 
 
 21 
 
 6 
 
 22 
 
 5 
 
 12 
 
 23 
 
 25 
 
 9 
 
 II 
 
 24 
 
 14 
 
 7 
 
 3 
 
 20 
 
 IO 
 
 19 
 
 15 
 
 13 
 
 20 
 
 4 
 
 18 
 
 17 
 
 16 
 
 2 
 
 8 
 
 z 
 
 21 
 
 22 
 
 7 
 
 6 
 
 23 
 
 S 
 
 12 
 
 9 
 
 I 4 
 
 II 
 
 24 
 
 25 
 
 3 
 
 21 
 
 15 
 
 8 
 
 10 
 
 20 
 
 13 
 
 z8 
 
 16 
 
 z 
 
 2 
 
 17 
 
 21 
 
 4 
 
 22 
 
 5 
 
 6 
 
 12 
 
 9 
 
 ii 
 
 19 
 
 23 
 
 14 
 
 25 
 
 3 
 
 7 
 
 24 
 
 22 
 
 10 
 
 4 
 
 17 
 
 20 
 
 IS 
 
 8 
 
 19 
 
 18 
 
 13 
 
 i 
 
 21 
 
 16 
 
 2 
 
 6 
 
 22 
 
 S 
 
 12 
 
 7 
 
 23 
 
 14 
 
 9 
 
 24 
 
 25 
 
 ii 
 
 3 
 
 23 
 
 IO 
 
 8 
 
 17 
 
 IS 
 
 4 
 
 2 
 
 13 
 
 16 
 
 19 
 
 20 
 
 Z 
 
 18 
 
 21 
 
 16 
 
 7 
 
 22 
 
 12 
 
 5 
 
 14 
 
 ii 
 
 9 
 
 25 
 
 23 
 
 3 
 
 24 
 
 34 
 
 10 
 
 4 
 
 IS 
 
 I 
 
 19 
 
 13 
 
 17 
 
 20 
 
 2 
 
 21 
 
 16 
 
 8 
 
 7 
 
 6 
 
 18 
 
 5 
 
 22 
 
 ii 
 
 14 
 
 24 
 
 12 
 
 9 
 
 23 
 
 25 
 
 3 
 
 itself. If one had table IV before him and were asked how the 
 readers ranked composition No. i he could see by consulting the 
 table, that reader: 
 
 No. i ranked composition No. i, i2th 
 
 No. 2 ranked the same composition nth 
 
 No. 3 " 7th 
 
 No. 4 " " " ioth 
 
 No. 5 " " " 6th 
 
 No. 6 " i2th 
 
 No. 7 " " 4 th, and so on. 
 
 To show these facts about each composition in the most conven- 
 ient form a distribution was prepared for each composition show- 
 ing the rank given that composition by each of the twenty-four 
 readers. Here is such a distribution for composition No. i. 
 
HARVARD-NEWTON BULLETIN 33 
 
 COMPOSITION NO. i 
 Rank 
 
 i 2 3 4 5 6 7 8 Q 10 ii 12 13 14 15 16 17 18 IQ 20 21 22 23 24 25 
 
 23 
 
 24 ............ 18 20 
 
 15 19 17 .. 9 12 21 .. 22 8 6 
 
 14 16 7 10 5 3 ii .. 4 2 i 13 
 
 The series of numbers at the top ranging from one to twenty- 
 five, represents the possible ranks, any one of which this composi- 
 tion might have been assigned by the readers. The figures near 
 the bottom are the numbers of the readers giving composition 
 No. i the ranks indicated. For example, composition No. i 
 was ranked second by readers Nos. 14 and 15; it was ranked 
 third by readers Nos. 16 and 19; it was ranked fourth by readers 
 Nos. 7, 17, and 24, and so on. Similar distributions were prepared 
 for each composition, but space does not permit of their intro- 
 duction in this monograph. They may be prepared from Table 
 IV by anyone interested. 
 
 These distributions show that readers agreed well as to the 
 rank of some compositions, while in other cases the greatest dif- 
 ference in judgment prevails. For example, composition No. 10 
 is ranked first by twenty-one out of twenty-four readers, and one 
 of these three gave it the highest grade, viz. 95 %. In the case of 
 composition No. 8, there seems to be little agreement. One 
 reader ranks it as high as ist, while one ranks it as low as i6th: the 
 remaining readers are scattered between these two extremes, with 
 a group of four readers ranking it in second place. Obviously, a 
 scattering distribution means a wide diversity of judgments in 
 ranking, and conversely, a heaping up of the figures under a few 
 ranks means a greater unanimity of judgments among readers. 
 The following table shows a summary of the distribution for each 
 composition. 
 
 Table V shows the distribution of the ranks given each com- 
 position in description. Compositions are arranged in serial 
 order according to the " Median Ranks," beginning with Median 
 Rank I. 
 
34 
 
 HARVARD-NEWTON BULLETIN 
 
 Comp. 
 No. 
 
 TABLE V 
 
 Rank 
 Median 
 
 10 
 IS 
 4 
 8 
 13 
 20 
 
 2 
 
 I 
 17 
 19 
 
 18 
 16 
 
 21 
 
 6 
 
 22 
 7 
 
 5 
 
 12 
 
 23 
 14 
 9 
 II 
 24 
 25 
 
 
 I.O 
 
 3-0 
 4-5 
 6.0 
 6.0 
 6.0 
 7-0 
 7-5 
 8.0 
 9-0 
 g.o 
 9-5 
 
 II.O 
 
 14.0 
 iS-o 
 16.5 
 17.0 
 17.0 
 I9-S 
 I9-S 
 
 2O.O 
 22.0 
 22.0 
 23-5 
 2S.O 
 
 
 I 8 7 3 ? II. 
 
 
 
 
 2I3433I2 2 I 
 
 
 
 
 
 
 . 31.2.215 23.1...!.. 
 
 
 2 
 
 ..2111331 43 ..' I !'..".. 
 
 1322 325 I. ..I.. 
 
 II 3.2 2221. I. 
 2. .-3 6II2.3I 
 1.2 4334.12 
 
 
 . i 3 i 6 3 I L 
 
 \ 4 
 I 2 4 
 Z I . 
 546l 
 
 I 22.1613 
 
 I 124166 
 
 I 4122 
 
 
 
 I . 5 3 I 
 
 Not only is there great diversity of judgments among readers, 
 but there are cases where some readers seem to be extremely 
 radical in their judgment of a composition on which other readers 
 agree. For example, one reader ranks composition No. 4, i6th, 
 while most of the other readers rank it from 8th to Qth, to 2d or 
 3d. Several readers seem likewise to be radical or extreme in 
 their judgment of the rank of composition No. 7. In view of 
 these observations we were led to adopt a method of segregating 
 that 25% of the readers who were radical or extreme in their 
 ranking. The lines were drawn in the distributions so as to cut 
 off the 25% or less of the readers who were considered extreme. 
 The maximum variation, both of the whole series and of the 
 central 75% of the cases, has also been computed for each com- 
 position. The following table shows these statistical facts. 
 
HARVARD-NEWTON BULLETIN 
 
 35 
 
 Table VI shows (a) the highest rank, (b) the lowest rank, (c) the 
 maximum variation in rank of the series, (d) the maximum varia- 
 tion of 75 % of the cases, (e) the median rank, (/) the average 
 deviation for each of the twenty-five compositions in description. 
 
 TABLE VI 
 
 Composition 
 Number 
 
 Highest 
 Rank 
 
 Lowest 
 Rank 
 
 Maximum 
 Variation in 
 Series 
 
 Maximum 
 Variation of 
 75% of Cases 
 
 Median 
 Rank 
 
 Average 
 Deviation 
 
 I 
 
 2 
 
 13 
 
 II 
 
 8 
 
 7-5 
 
 3-17 
 
 2 
 
 5 
 
 15 
 
 10 
 
 4 
 
 7.0 
 
 2.17 
 
 3 
 
 21 
 
 25 
 
 4 
 
 2 
 
 25-0 
 
 71 
 
 4 
 
 2 
 
 16 
 
 14 
 
 6 
 
 4-5 
 
 2.8 3 
 
 5 
 
 14 
 
 21 
 
 7 
 
 5 
 
 17.0 
 
 1.46 
 
 6 
 
 8 
 
 20 
 
 12 
 
 6 
 
 14.0 
 
 2.00 
 
 7 
 
 IO 
 
 24 
 
 14 
 
 7 
 
 16.5 
 
 2.67 
 
 8 
 
 I 
 
 16 
 
 IS 
 
 10 
 
 6.0 
 
 4.04 
 
 9 
 
 14 
 
 23 
 
 9 
 
 4 
 
 20.0 
 
 I. S 8 
 
 10 
 
 I 
 
 3 
 
 2 
 
 
 
 I.O 
 
 25 
 
 ii 
 
 16 
 
 25 
 
 9 
 
 S 
 
 22.O 
 
 2.08 
 
 12 
 
 14 
 
 23 
 
 9 
 
 4 
 
 17.0 
 
 I.7I 
 
 13 
 
 i 
 
 12 
 
 ii 
 
 7 
 
 6.0 
 
 2.50 
 
 14 
 
 9 
 
 24 
 
 IS 
 
 6 
 
 19-5 
 
 2-75 
 
 15 
 
 i 
 
 10 
 
 9 
 
 3 
 
 3-0 
 
 i-33 
 
 16 
 
 3 
 
 17 
 
 14 
 
 7 
 
 9-5 
 
 3-25 
 
 17 
 
 3 
 
 14 
 
 ii 
 
 7 
 
 8.0 
 
 2.58 
 
 18 
 
 5 
 
 15 
 
 IO 
 
 6 
 
 9-5 
 
 2.46 
 
 19 
 
 2 
 
 19 
 
 17 
 
 8 
 
 9.0 
 
 3-25 
 
 20 
 
 2 
 
 16 
 
 14 
 
 7 
 
 6.0 
 
 3-oo 
 
 21 
 
 6 
 
 19 
 
 13 
 
 6 
 
 II.O 
 
 2.38 
 
 22 
 
 8 
 
 21 
 
 13 
 
 7 
 
 15.0 
 
 2.46 
 
 23 
 
 14 
 
 23 
 
 9 
 
 4 
 
 !9-5 
 
 2.00 
 
 24 
 
 17 
 
 25 
 
 8 
 
 4 
 
 22.0 
 
 1.38 
 
 25 
 
 19 
 
 25 
 
 6 
 
 3 
 
 23-5 
 
 I.2Q 
 
36 HARVARD-NEWTON BULLETIN 
 
 SELECTING THE SCALE 
 (a) THE BASIS 
 
 In the preceding pages, the following statistical information 
 has been worked out as a basis for selecting the six compositions 
 for the proposed scale: 
 
 A. From the grades given each composition by the twenty-four 
 readers. 
 
 (a) The maximum variation in grades. 
 
 (b) The mean or average grade. 
 
 (c) The median grade. 
 
 B. From the rank or relative position assigned each composi- 
 tion by the twenty-four readers. 
 
 (a) The maximum variation of the whole series. 
 
 (b) The extreme variation of 75% of the most central cases. 
 
 (c) The median rank. 
 
 (d) The average deviation. 
 
 (b) FIXING STANDARD REQUIREMENTS 
 
 Before selecting the scale on the basis of this information it is 
 necessary to answer these questions. What are the standard or 
 ideal requirements for each of the six compositions of the pro- 
 posed scale in terms of the suggested basis of selection ? What 
 ought the maximum variation, the mean, and the median to be, 
 according to the grade basis of selection ? What ought the two 
 maximum variations, the median rank, and the average deviation 
 to be, according to the rank basis of selection ? We shall now 
 proceed to answer these questions, so far as we can. 
 
 MAXIMUM VARIATIONS AND AVERAGE DEVIATIONS 
 
 Large maximum variation, in either the grades or ranks, 
 indicates great diversity of judgment among the readers, and 
 conversely small maximum variation indicates better agreement 
 in judgment among readers. The former has been shown in the 
 case of composition No. 8, and the latter in the case of composi- 
 tion No. 10 (See p. 34). Obviously, although no definite fixed 
 
HARVARD-NEWTON BULLETIN 37 
 
 standard for variations can be established for any of the six 
 compositions of the proposed scale, it will be agreed that, other 
 things being equal, that composition is most suitable for a place 
 in the scale about which there is greatest agreement in judgment 
 both in grade and rank, as shown by a small maximum variation 
 and a small average deviation. 
 
 THE MEAN, OR AVERAGE GRADE 
 
 The standard requirements of the mean or average grade for 
 each of the six compositions of the scale is not difficult to deter- 
 mine. Our object in this whole study is to secure a scale of six 
 compositions which shall be as nearly as possible, typical, 95 %, 
 85%, 75%. 65%, 55%, and 45% compositions. Since our 
 ultimate object is expressed in these figures, the six compositions 
 selected ought to have a mean or average grade of 95%, 85%, 
 75 %> 6 5%> 55 %, and 45% respectively. 
 
 THE MEDIAN GRADE 
 
 \Vhat has been said about the mean or average grade applies 
 also to the median grade. The six compositions for the scale 
 should have respectively a median grade of 95 %, 85 %, 75 %, 65 %, 
 55%, and 45%. 
 
 THE MEDIAN RANK 
 
 The standard requirements for the median rank require some 
 explanation. It is important that the six compositions of the 
 scale should be approximately the same distance apart in rank. 
 A difference of 10% divides them in grades: how many points in 
 rank should divide them ? There are twenty-five ranks, to any 
 one of which a composition may be assigned by any reader. In 
 view of the fact that there are to be six compositions in the scale, 
 there must necessarily be five spaces or steps in the series. The 
 problem then, is to divide the total points in rank between the 
 first and last compositions in the series into five equal divisions 
 so that the six compositions selected for the scale will be equidis- 
 tant one from the other. 
 
38 HARVARD-NEWTON BULLETIN 
 
 It is obvious that beginning with the A or 95 % composition, 
 that composition should have a median rank of i.o. Deducting 
 this one from the twenty-five ranks, the total number, we have 
 twenty-four points of rank remaining to be divided into five equal 
 divisions. This gives 4.8 points in rank for each of the spaces 
 between compositions. By adding this amount 4.8, to i.o, the 
 median rank of the A or 95% composition, we have 5.8, the 
 median rank of the B or 85 % composition of the scale. By add- 
 ing 4.8 to each succeeding median rank we find the following 
 median rank for each of the six compositions. 
 
 Median Rank 
 
 A or 95 % composition i.o 
 
 B or 85 % " 5.8 
 
 C or 75 % 10.6 
 
 Dor 65% 15.4 
 
 E or 55 % 20.2 
 
 F or 45 % . . . 25.0 
 
 Because the ranks in the graphs are given only in whole num- 
 bers or in whole numbers and halves, it is clear that the decimals 
 in the above median rank numbers cannot be exactly obtained. 
 However, the median rank for each of the compositions in the 
 scale should be as near the above standard median ranks as 
 possible. 
 
 SUMMARY 
 
 From the above discussion, we may summarize the standard 
 requirements for each of the six compositions of the proposed 
 scale. As has been already stated, there is no fixed standard 
 requirement in the case of maximum variations, or average 
 deviation, either in grade or in rank. The only principle to fol- 
 low is this : the narrower the maximum variation the better. 
 
 The other items are summarized as follows: 
 
 Grade or Rank of Mean or Median Median 
 
 Composition in Scale Average Grade Rank 
 
 A Grade or 95 % 95 % 95 % i.o 
 
 B Grade or 85% 85% 85% 5.8 
 
 C Grade or 75 % 75 % 75 % 10.6 
 
 D Grade or 65 % 65 % 65 % 15.4 
 
 E Grade or 55 % 55 % 55 % 20.2 
 
 F Grade or 45 % 45 % 45 % 25.0 
 
HARVARD-NEWTON BULLETIN 39 
 
 (c) MAKING THE SELECTION 
 
 The first step in the selection of the six compositions to con- 
 stitute the scale was to classify the twenty-five compositions into 
 six groups corresponding to the six grades or ranks of composi- 
 tions in the proposed scale. Each composition was placed in 
 that group with which it was most closely identified according to 
 the several statistical items in the table. Such a classification 
 results as follows: 
 
 In general Composition Nos. 
 
 A group of compositions (All from 90% up) 10-15. 
 
 B 
 
 C " 
 
 D 
 
 E 
 
 F 
 
 ( " 80 to 89%) 1-2-4-8-13-17-19-20. 
 
 ( a a 70 a 79%) 16-18-21. 
 
 ( a 60 69%) 5-6-7-12-22. 
 
 ( 50 a 59%) 9-14-23. 
 
 ( " " 40 " 49%) 3-11-24-25. 
 
 Having thus classified the twenty-five compositions, the next 
 step was to select the one from each group which was shown by the 
 standard requirements to be most suitable for the scale. Begin- 
 ning with the A or 95 % composition, we shall consider each group 
 in order. In tabular form are given the numbers of the composi- 
 tions which are to be considered for that particular place in the 
 scale, together with all the statistical data pertaining to each 
 composition. Preceding the table in each case are given the stan- 
 dard requirements for the composition as worked out in the 
 preceding pages. Following the table is the discussion leading 
 to the selection of the most satisfactory composition. 
 
 " A " GRADE OR 95 % COMPOSITION 
 
 Standard Requirements: 
 
 Mean or average grade 95 % 
 
 Median grade 95 % 
 
 Median rank.. . i.o 
 
40 HARVARD-NEWTON BULLETIN 
 
 TABLE VII 
 EXPLANATION 
 
 Letters in the following table designate the following topics: 
 (a) The number of the composition; (6) The highest grade given it; (c) The 
 lowest grade given it; (d) The maximum variation in grade; (e) The mean or aver- 
 age grade; (/) The median grade; (g) The highest rank given it; (K) The lowest 
 rank given it; (i) The maximum variation in rank; (;') The maximum variation 
 in rank of 75% of the central cases; (k) The median rank; and (/) Average de- 
 viation. 
 
 abed e f g h i j k I 
 
 10 * 95 90 5 94.6 95 ist 3d 2 o i.o .25 
 15 95 80 15 89.6 90 ist loth 9 3 3.0 1.33 
 
 It is not difficult to select the better of these two compositions. 
 No reader felt that composition No. 10 was lower than 90% in 
 grade, or than 3d in rank, whereas No. 15 was graded as low as 
 80%, and ranked as low as loth among the twenty-five composi- 
 tions. The extreme variations are therefore, greater in the 
 case of composition No. 15. Obviously No. 10 was considered by 
 the readers to be more nearly the best composition, hence it is 
 selected as the "A" grade or 95% composition in the scale. 
 No. 15 is too good a composition to be considered for the next 
 composition in the scale, hence it has to be dropped from further 
 consideration. 
 
 " B " GRADE OR 85 % COMPOSITION 
 
 Standard Requirements: 
 
 Mean or Average Grade 85 % 
 
 Median Grade 85 % 
 
 Median Rank 5.8 
 
 From the table below, one sees at once that, judged by these 
 statistical items, several of these compositions are not really 
 competitors for the second place in the scale. Nos. i and 2 are 
 too poor, i. e., they are graded and ranked too low. No. 4 has in 
 its favor the mean and median grades of 84.3% and 85.5% 
 respectively, but it has a wide extreme variation in both grade 
 and rank. Also, its median rank is only 4.5, whereas our require- 
 
 1 Selected for composition No. i in the scale. 
 
HARVARD-NEWTON BULLETIN 41 
 
 TABLE VIII 
 EXPLANATION 
 
 Letters in the following table designate the following topics: 
 (a) The number of the composition; (b) The highest grade given it; (c) The 
 lowest grade give it; (d) The maximum variation in grade; (e) The mean or 
 average grade ; (/) The median grade ; (g) The highest rank given it ; (K) The 
 lowest rank given it ; (i) The maximum variation in rank ; (j) The maximum varia- 
 tion in rank of 75 % of the central cases ; (k) The median rank, and (/) The average 
 deviation. 
 
 a 
 
 b 
 
 c 
 
 d 
 
 e 
 
 / 
 
 g 
 
 h 
 
 i 
 
 j 
 
 k 
 
 / 
 
 i 
 
 95 
 
 68 
 
 27 
 
 81.9 
 
 83.0 
 
 2 d 
 
 i3th 
 
 ii 
 
 8 
 
 7-5 
 
 3-i7 
 
 2 
 
 90 
 
 64 
 
 26 
 
 80.0 
 
 80.0 
 
 5th 
 
 iSth 
 
 10 
 
 4 
 
 7.0 
 
 2.17 
 
 4 
 
 94 
 
 63 
 
 3i 
 
 84-3 
 
 85-5 
 
 2d 
 
 i6th 
 
 14 
 
 6 
 
 4-5 
 
 2.83 
 
 8 
 
 95 
 
 52 
 
 43 
 
 82.3 
 
 85.0 
 
 ISt 
 
 i6th 
 
 15 
 
 10 
 
 6.0 
 
 4.04 
 
 I3 1 
 
 95 
 
 ' 7i 
 
 24 
 
 83-7 
 
 85.0 
 
 ISt 
 
 i2th 
 
 ii 
 
 7 
 
 6.0 
 
 2.50 
 
 i? 
 
 93 
 
 63 
 
 30 
 
 81.0 
 
 81.5 
 
 3d 
 
 i4th 
 
 ii 
 
 7 
 
 8.0 
 
 2.58 
 
 iQ 
 
 92 
 
 60 
 
 32 
 
 79.6 
 
 80.0 
 
 2d 
 
 i9th 
 
 17 
 
 8 
 
 9.0 
 
 3-25 
 
 20 
 
 92 
 
 70 
 
 22 
 
 82.7 
 
 85.0 
 
 2d 
 
 1 6th 
 
 14 
 
 7 
 
 6.0 
 
 3.00 
 
 ments call for a median rank of 5.8. No. 8 is one of those com- 
 positions on which teachers cannot agree. Its extreme variation 
 puts it beyond our further consideration here. Comparatively, 
 No. 13 has not a wide variation in grade or rank, and it has a 
 mean and median grade of 83.7 and 85.0 respectively. It is clear 
 that this composition is a strong competitor for the second place. 
 No. 17 has a wider variation than No. 13, and its mean and 
 medians indicate that it is too poor for an 85% composition. 
 No. 19 is likewise too poor. No. 20 has a narrow extreme varia- 
 tion, its mean is a little low, but its median is perfect. Its ex- 
 treme variation by rank is comparatively wide, while its median 
 rank is satisfactory. Clearly the selection of an 85 % composi- 
 tion for this scale must be made between Nos. 13 and 20. We 
 select No. 13 because (a) of its slightly narrower extreme varia- 
 tion by grade, because (b) of its mean of 83.7 which is i % nearer 
 the standard than the mean of No. 20, because (c) its extreme 
 variation in rank is three points narrower than that of No. 20 and 
 because, (d) in other respects it is as good as No. 20. 
 
 1 Selected for Composition No. 2 in the scale. 
 
42 HARVARD-NEWTON BULLETIN 
 
 " C " GRADE OR 75 % COMPOSITION 
 
 Standard Requirements: 
 
 Mean or Average Grade 75 % 
 
 Median Grade 75 % 
 
 Median Rank . . 10.6 
 
 TABLE IX 
 EXPLANATION 
 
 Letters in the following table designate the following topics: 
 (a) The number of the composition; (b) The highest grade given it; (c) The 
 lowest grade given it; (d) The maximum variation in grade; (e) The mean or 
 average grade; (/) The median grade; (g) The highest rank given it; (h) The 
 lowest rank given it; (i) The maximum variation in rank; (/) The maximum 
 variation in rank of 75 % of the central cases; (k) The median rank, and (/) The 
 average deviation. 
 
 abed e f g h i j k I 
 
 16 92 68 24 78.2 78.5 3d i7th 14 7 9.5 3.25 
 
 18 90 60 30 79.9 75.0 2d i9th 17 8 9.5 2.46 
 
 21 l 89 54 35 76.1 77.0 6th i9th 13 6 n.o 2.38 
 
 Obviously compositions Nos. 16 and 18 are too good composi- 
 tions for typical 75 % grade compositions. The mean or average 
 grade of the former is 78.2 and that of the latter is 79.9. In rank 
 likewise they are better than No. 21 and too good for our purpose. 
 That leaves only No. 21 to be selected for the C grade or 75% 
 composition. Even though this is the best selection that can be 
 made under the circumstances, one could wish that the extreme 
 variations were less and that the other statistical items were 
 nearer the standard requirements. 
 
 " D " GRADE OR 65 % COMPOSITION 
 
 Standard Requirements: 
 
 Mean or Average Grade 65 % 
 
 Median Grade 65 % 
 
 Median Rank 15.4 
 
 1 Selected for Composition No. 3 in the scale. 
 
HARVARD-NEWTON BULLETIN 43 
 
 TABLE X 
 EXPLANATION 
 
 Letters in the following table designate the following topics: 
 (a) The number of the composition; (b) The highest grade given it; (c) The 
 lowest grade given it; (d) The maximum variation in grade; (e) The mean or 
 average grade; (/) The median grade; (g) The highest rank given it; (h) The 
 lowest rank given it; (i) The maximum variation in rank; (j) The maximum 
 variation in rank of 75 % of the most central cases; (k) The median rank, and (/) 
 The average deviation. 
 
 abode f g h i j k I 
 
 5 78 50 28 61.1 60.0 i4th 2ist 7 5 17.0 1.46 
 
 6 88 50 38 69.4 69.5 8th 2oth 12 6 14.0 2.0 
 
 7 80 40 40 63.5 65.0 loth 24th 14 7 16.5 2.67 
 12 75 42 33 59-9 60.0 i4th 23d 9 4 17.0 1.71 
 22 : 86 47 39 66.6 66.5 8th 2ist 13 7 15.0 2.46 
 
 Clearly Nos. 5 and 12 are too poor for a typical 65% composi- 
 tion. No. 7 has a wide extreme variation in both grade and 
 rank: although its mean and median grades might do, its median 
 rank is too low. Hence No. 22 is selected, even though its mean 
 and median grades are slightly high to meet the standard require- 
 ments. 
 
 " E " GRADE OR 55 % COMPOSITION 
 Standard Requirements: 
 
 Mean or Average Grade 55 % 
 
 Median Grade 55 % 
 
 Median Rank 20.2 
 
 TABLE XI 
 EXPLANATION 
 
 Letters in the following table designate the following topics: 
 (a) The number of the composition; (b) The highest grade given it; (c) The 
 lowest grade given it; (d) The maximum variation in grade; (e) The mean or 
 average grade; (/) The median grade; (g) The highest rank given it; (h) The 
 lowest rank given it; (*) The maximum variation in rank; (j) The maximum 
 variation in rank of 75 % of the most central cases; (k) The median rank, and (/) 
 The average deviation. 
 
 a 
 
 b 
 
 c 
 
 d 
 
 e 
 
 / 
 
 g 
 
 i 
 
 i 
 
 J 
 
 i 
 
 | 
 
 9 
 
 75 
 
 40 
 
 35 
 
 56.1 
 
 58.5 
 
 i4th 
 
 2 3 d 
 
 9 
 
 4 
 
 2O.O 
 
 1.58 
 
 14 
 
 76 
 
 40 
 
 36 
 
 55-4 
 
 53-5 
 
 9th 
 
 24th 
 
 15 
 
 6 
 
 19-5 
 
 2-75 
 
 2 3 2 
 
 74 
 
 40 
 
 34 
 
 55-4 
 
 57-5 
 
 i4th 
 
 23d 
 
 9 
 
 4 
 
 19-5 
 
 2.00 
 
 1 Selected for Composition No. 4 in the scale. 
 
 2 Selected as Composition No. 5 in the scale. 
 
44 HARVARD-NEWTON BULLETIN 
 
 In considering No. 9 it is clear that it is a little too good for a 
 typical 55% composition, as shown in two of the three items of 
 the standard requirements. The item of median rank is perfect. 
 However, it should be pointed out that the median rank of the 
 composition selected to occupy a place immediately preceding 
 this was 15.0 instead of the standard requirement of 15.4. Hence, 
 the median rank of composition 55% in the scale should be 
 slightly under 20.0 rather than over it. Between Nos. 14 and 23 
 the advantage of No. 23 is in the narrower extreme variation 
 by rank and in the slightly higher median grade. Hence, No. 23 
 is selected. 
 
 " F " GRADE OR 45 % COMPOSITION 
 
 Standard Requirements: 
 
 Mean or Average Grade 45 % 
 
 Median Grade 45 % 
 
 Median Rank 25.0 
 
 TABLE XII 
 
 EXPLANATION 
 
 Letters in the following table designate the following topics: 
 (a) The number of the composition; (6) The highest grade given it; (c) The 
 lowest grade given it; (d) The maximum variation in grade; () The mean or 
 average grade; (/) The median grade; (g) The highest rank given it; (h} The 
 lowest rank given it; (i) The maximum variation in rank; (j) The maximum 
 variation in rank of 75 % of the most central cases; (k) The median rank, and (/) 
 The average deviation. 
 
 a b c d e f g h i j k I 
 
 3 50 30 20 42.7 41.0 2ist 25th 4 2 25.0 .71 
 
 ii 65 40 25 49.5 49.5 i6th 25th 9 5 22.0 2.08 
 
 24 73 30 42 48.9 48.0 i7th 25th 8 4 22.0 1.38 
 
 25 * 62 20 42 44.9 45.0 i9th 25th 6 3 23.5 1.29 
 
 The objection to No. 3 is that, although all agree that it is 25th 
 composition in rank, or the poorest in the series, the grades given 
 it indicate that it is too poor. On the other hand, No. 1 1 is too 
 good. The choice then lies between Nos. 24 and 25. It is clear 
 that No. 25 more nearly meets the standard requirements. No. 
 24 is a little too good. It should be pointed out that the median 
 
 , * Selected as Composition No. 6 in the scale. 
 
HARVARD-NEWTON BULLETIN 
 
 45 
 
 rank of composition No. 25 is really more satisfactory than at 
 first appears. The median rank of the 55% composition was 
 19.5, a little higher than the standard requirements: if to this is 
 added 4.8, the difference in points of rank between compositions, 
 we have 24.3, a little less than the standard requirements indicate. 
 Hence, No. 25 is selected for the sixth and last composition in the 
 scale. 
 
 SUMMARY 
 
 The following table shows at a glance how the statistical items 
 pertaining to each composition selected for the scale, 1 compare 
 with the standard requirements established for each composition. 
 
 
 
 
 Mean 
 
 
 Median 
 
 
 Median 
 
 
 No. of 
 
 Standard 
 
 Grade of 
 
 Standard 
 
 Grade of 
 
 Standard 
 
 Rank of 
 
 Compositions of the Scale 
 
 Compo- 
 sition 
 Selected 
 
 Mean or 
 Average 
 Grade 
 
 Compo- 
 sition 
 Selected 
 
 Median 
 Grade 
 
 Compo- 
 sition 
 Selected 
 
 Median 
 Rank 
 
 Compo- 
 sition 
 Selected 
 
 A grade or 95 % . . 
 
 IO 
 
 QC 
 
 04-6 
 
 QC 
 
 or O 
 
 I.O 
 
 I O 
 
 B 85% 
 
 13 
 
 8 5 
 
 83.5 
 
 8 5 
 
 85.0 
 
 5-8 
 
 6.0 
 
 C " 75% 
 
 21 
 
 75 
 
 76.1 
 
 75 
 
 77.0 
 
 10.6 
 
 II.O 
 
 D 65% 
 
 22 
 
 65 
 
 66.6 
 
 65 
 
 66. 5 
 
 15-4 
 
 15-0 
 
 E 55% 
 
 23 
 
 55 
 
 55-4 
 
 55 
 
 57-5 
 
 20. 2 
 
 iQ-5 
 
 F 45% 
 
 25 
 
 45 
 
 44.9 
 
 45 
 
 45 - 
 
 25.0 
 
 23-5 
 
 The chart on page 46 indicates the distribution of ranks for each 
 of the six compositions constituting the description scale. The 
 median is indicated by the line dividing the distribution. For an 
 explanation of this chart see pages 32 and 33. 
 
 DEFINING THE MERITS AND DEFECTS OF THE COMPOSI- 
 TIONS IN THE SCALES 
 
 After the scales had been selected, the next step was to define 
 the merits and defects of each composition in each scale. This 
 was done by a committee consisting of high school and elementary 
 school teachers with the cooperation of the teachers and principals 
 who had originally read the compositions. 
 
 1 For the Compositions of the scale, see pp. 49-57. 
 
4 6 
 
 HARVARD-NEWTON BULLETIN 
 
 
 Nc 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 R, 
 
 IN 
 
 < 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Nc 
 
 c 
 
 
 Jo 
 
 rip 
 
 OS 
 
 tic 
 
 n 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 :o 
 
 "P 
 
 OS 
 
 ticn 
 
 
 S 
 
 :al 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 c 
 
 or 
 
 P 1 
 
 et( 
 
 
 
 
 
 
 
 24 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 < 
 
 et 
 
 
 
 
 
 
 
 
 23 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 22 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 20 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 19 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 18 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 17 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 16 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 15 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 13 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Nc 
 
 . 1 
 
 
 
 
 
 12 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 N 
 
 0. 
 
 10 
 
 
 
 
 
 
 
 11 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 10 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 9 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 8 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 7 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 5 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 4 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 3 
 
 
 21 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 2 
 
 
 14 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 1 
 
 
 6 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 21 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 N 
 
 . ! 
 
 
 
 
 
 
 
 
 20 
 
 19 
 
 24 
 
 
 
 
 18 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 fv 
 
 0. 
 
 13 
 
 
 
 
 
 
 
 
 
 17 
 
 11 
 
 8 
 
 13 
 
 23 
 
 
 22 
 
 10 
 
 
 16 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 6 
 
 
 12 
 
 1 
 
 2 
 
 3 
 
 15 
 
 4 
 
 9 
 
 5 
 
 14 
 
 7 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 .23 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 20 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 N< 
 
 
 
 
 
 
 
 
 
 
 
 
 13 
 
 
 
 24 
 
 22 
 
 
 19 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 No. 
 
 21 
 
 
 
 
 
 
 
 
 
 
 
 
 
 10 
 
 18 
 
 8 
 
 2 
 
 21 
 
 12 
 
 15 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 17 
 
 9 
 
 16 
 
 7 
 
 1 
 
 6 
 
 11 
 
 14 
 
 4 
 
 5 
 
 
 
 
 3 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 22 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 N 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 19 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 No. 
 
 22 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 20 
 
 11 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 21 
 
 17 
 
 8 
 
 
 
 
 
 12 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 10 
 
 
 
 
 
 16 
 
 15 
 
 7 
 
 
 
 14 
 
 
 5 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 9 
 
 
 13 
 
 
 
 3 
 
 2 
 
 4 
 
 23 
 
 24 
 
 6 
 
 
 1 
 
 18 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 22 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 18 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 16 
 
 
 
 17 
 
 24 
 
 
 
 
 
 
 
 
 
 
 BJ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 20 
 
 
 13 
 
 21 
 
 
 14 
 
 23 
 
 
 
 
 
 
 
 N 
 
 0. 
 
 23 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 10 
 
 
 6 
 
 11 
 
 
 5 
 
 15 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 8 
 
 
 9 
 
 3 
 
 19 
 
 2 
 
 4 
 
 7 
 
 1 
 
 12 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 24 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 20 
 
 
 
 
 
 
 
 
 
 N 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 23 
 
 
 17 
 
 18 
 
 
 
 
 
 N 
 
 0. 
 
 25 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 21 
 
 22 
 
 15 
 
 14 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 16 
 
 11 
 
 13 
 
 10 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 6 
 
 7 
 
 12 
 
 6 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 19 
 
 8 
 
 9 
 
 3 
 
 4 
 
 2 
 
 1 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
HARVARD-NEWTON BULLETIN 47 
 
 The six compositions of the description scale formed the basis 
 of a teachers' meeting, consisting of eighth grade teachers, ele- 
 mentary school principals, and high school teachers of English. 
 In preparation for the meeting each person was asked to define, in 
 a succinct paragraph, the merits and defects of one of the six 
 compositions in the scale. This assignment had two results: 
 first, it prepared the teachers to participate in the discussion, and 
 second, it furnished five or six paragraphs from as many different 
 people setting forth the merits and defects of each composition 
 as each reader saw them. In further preparation, two persons 
 were asked to lead the discussion on each composition. With this 
 preliminary preparation, the meeting proved exceptionally inter- 
 esting and profitable to the teachers. It was interesting because 
 it " moved," and it was profitable because it was devoted to the 
 consideration of some practical everyday school matters on which 
 every teacher was glad to obtain suggestions. 
 
 At the close of the meeting the written paragraphs mentioned 
 above, were collected and turned over to the committee which was 
 to prepare the necessary paragraphs. The preparation of these 
 paragraphs was not considered an easy task and the committee 
 did its work conscientiously and with thoroughness. After 
 tentative paragraphs had been prepared by each member of the 
 committee according to agreement the committee held a meeting 
 to consider them. Besides the five members of the committee 
 Superintendent Spaulding, Professor Holmes of Harvard, and 
 the Joseph Lee Fellow attended the meeting. The paragraphs 
 were read and discussed at length. It was agreed that the par- 
 agraph accompanying each composition should state, in general, 
 why each composition was better than the one below it in the scale : 
 why it was poorer than the one above it in the scale, and also its 
 own intrinsic worth. With the suggestions from the discussion 
 of the committee, the committee revised the paragraphs and 
 prepared the final definition of merits and defects as they are 
 found accompanying each composition in the description scale. 
 
 Through the courtesy of Mr. Charles S. Thomas, Head of the 
 Department of English in the Classical High School, Newton, and 
 
48 HARVARD-NEWTON BULLETIN 
 
 Editor of the English Leaflets of the New England Association of 
 Teachers of English, the Description Scale and the accompanying 
 paragraphs were published in the February issue of the Leaflet 
 with the request that members of the association offer suggestions 
 for improving them before final publication. 
 
 The same general procedure was followed in Exposition, Argu- 
 mentation, and Narration, except that the teachers were divided 
 into two groups, and each group considered two of the above 
 types. In the case of these scales, also, the final paragraphs were 
 not prepared in time to be presented in the English Teachers 
 Leaflet to obtain the criticism of the members of the association. 
 
HARVARD-NEWTON BULLETIN 33 
 
 COMPOSITION NO. i 
 
 Rank 
 
 i 2 3 4 5 6 7 8 g 10 ii 12 13 14 15 16 17 18 IQ 20 21 22 23 24 
 
 23 
 
 24 18 20 
 
 .. 15 19 17 .. 9 12 21 .. 22 8 6 
 
 . . 14 16 7 10 5 3 ii . . 4 2 i 13 
 
 The series of numbers at the top ranging from one to twenty- 
 five, represents the possible ranks, any one of which this composi- 
 tion might have been assigned by the readers. The figures near 
 the bottom are the numbers of the readers giving composition 
 No. i the ranks indicated. For example, composition No. i 
 was ranked second by readers Nos. 14 and 15; it was ranked 
 third by readers Nos. 16 and 19; it was ranked fourth by readers 
 Nos. 7, 17, and 24, and so on. Similar distributions were prepared 
 for each composition, but space does not permit of their intro- 
 duction in this monograph. They may be prepared from Table 
 IV by anyone interested. 
 
 These distributions show that readers agreed well as to the 
 rank of some compositions, while in other cases the greatest dif- 
 ference in judgment prevails. For example, composition No. 10 
 is ranked first by twenty-one out of twenty-four readers, and one 
 of these three gave it the highest grade, viz. 95 %. In the case of 
 composition No. 8, there seems to be little agreement. One 
 reader ranks it as high as ist, while one ranks it as low as i6th : the 
 remaining readers are scattered between these two extremes, with 
 a group of four readers ranking it in second place. Obviously, a 
 scattering distribution means a wide diversity of judgments in 
 ranking, and conversely, a heaping up of the figures under a few 
 ranks means a greater unanimity of judgments among readers. 
 The following table shows a summary of the distribution for each 
 composition. 
 
 Table V shows the distribution of the ranks given each com- 
 position in description. Compositions are arranged in serial 
 order according to the " Median Ranks,' 7 beginning with Median 
 Rank I. 
 
34 
 
 HARVARD-NEWTON BULLETIN 
 
 TABLE V 
 
 Rank 
 
 Comp. Median 
 No. Rank / a 3 4 5 6 7 8 9 10 n 12 13 14 15 16 17 18 19 20 21 22 23 24 25 
 
 IO 
 
 I O 
 
 zi 3 . 
 
 
 
 
 
 15 
 
 3-O 
 
 18733 . . i 
 
 r 
 
 
 
 
 4 
 8 
 13 
 
 4-5 
 6.0 
 60 
 
 . 51622121 
 I433-3II. 
 i 2343212 
 
 
 . i 
 
 2 . 
 2 
 
 . I .... 
 [21.... 
 
 . . . 
 
 20 
 
 2 
 
 6.0 
 7-0 
 
 .21343-31 
 ....17513 
 
 
 2 
 I I 
 
 . I .... 
 II 
 
 . . . 
 
 
 80 
 
 
 
 
 
 
 
 
 31 2 215 
 
 
 2 3 
 
 I i 
 
 
 18 
 
 9 
 
 31233 
 
 
 6 i 
 
 I 
 
 
 16 
 
 9-S 
 
 2III33I 
 
 
 4 3 
 
 I ... 
 
 
 21 
 
 II.O 
 
 1322 
 
 * 
 
 [ 2 5 
 
 I i 
 
 
 6 
 
 22 
 
 14.0 
 15 o 
 
 . II 
 
 . 2 
 
 : 
 
 t 2 
 
 2 2 3 I . I 
 
 6 I i 2 3 
 
 
 7 
 
 16 5 
 
 
 i 
 
 2 
 
 4334 i 
 
 ' 
 
 
 
 
 
 
 
 
 12 
 
 17.0 
 
 
 
 
 156421 
 
 2 
 
 21 
 
 
 
 
 
 
 
 14 
 
 10 S 
 
 I 
 
 
 
 22 I 6 I 
 
 
 g 
 
 20 o 
 
 
 
 3 
 
 [ 12416 
 
 
 II 
 
 22. 
 
 
 
 
 I 412 
 
 ? 6 i 
 
 24 
 
 22. 
 
 
 
 
 I 13 
 
 8 23 
 
 25 
 
 23.5 
 
 
 
 
 i i 
 
 5 75 
 
 3 
 
 2.S.O 
 
 
 
 
 
 
 Not only is there great diversity of judgments among readers, 
 but there are cases where some readers seem to be extremely 
 radical in their judgment of a composition on which other readers 
 agree. For example, one reader ranks composition No. 4, i6th, 
 while most of the other readers rank it from 8th to gth, to 2d or 
 3d. Several readers seem likewise to be radical or extreme in 
 their judgment of the rank of composition No. 7. In view of 
 these observations we were led to adopt a method of segregating 
 that 25% of the readers who were radical or extreme in their 
 ranking. The lines were drawn in the distributions so as to cut 
 off the 25% or less of the readers who were considered extreme. 
 The maximum variation, both of the whole series and of the 
 central 75% of the cases, has also been computed for each com- 
 position. The following table shows these statistical facts. 
 
HARVARD-NEWTON BULLETIN 
 
 35 
 
 Table VI shows (a) the highest rank, (b) the lowest rank, (c) the 
 maximum variation in rank of the series, (d) the maximum varia- 
 tion of 75% of the cases, (e) the median rank, (/) the average 
 deviation for each of the twenty-five compositions in description. 
 
 TABLE VI 
 
 Composition 
 Number 
 
 Highest 
 Rank 
 
 Lowest 
 Rank 
 
 Maximum 
 Variation in 
 Series 
 
 Maximum 
 Variation of 
 75% of Cases 
 
 Median 
 Rank 
 
 Average 
 Deviation 
 
 I 
 
 2 
 
 13 
 
 II 
 
 8 
 
 7-5 
 
 3-17 
 
 2 
 
 5 
 
 15 
 
 IO 
 
 4 
 
 7.0 
 
 2.17 
 
 3 
 
 21 
 
 25 
 
 4 
 
 2 
 
 25.0 
 
 71 
 
 4 
 
 2 
 
 16 
 
 14 
 
 6 
 
 4-5 
 
 2.8 3 
 
 5 
 
 14 
 
 21 
 
 7 
 
 5 
 
 17.0 
 
 1.46 
 
 6 
 
 8 
 
 2O 
 
 12 
 
 6 
 
 14.0 
 
 2.OO 
 
 7 
 
 10 
 
 24 
 
 14 
 
 7 
 
 I6. S 
 
 2.6 7 
 
 8 
 
 I 
 
 16 
 
 15 
 
 IO 
 
 6.0 
 
 4.04 
 
 9 
 
 14 
 
 23 
 
 9 
 
 4 
 
 20.0 
 
 1.58 
 
 10 
 
 I 
 
 3 
 
 2 
 
 
 
 I.O 
 
 25 
 
 ii 
 
 16 
 
 25 
 
 9 
 
 5 
 
 22.0 
 
 2.08 
 
 12 
 
 14 
 
 23 
 
 9 
 
 4 
 
 17.0 
 
 I.7I 
 
 13 
 
 i 
 
 12 
 
 ii 
 
 7 
 
 6.0 
 
 2.50 
 
 14 
 
 9 
 
 24 
 
 15 
 
 6 
 
 19-5 
 
 2-75 
 
 15 
 
 i 
 
 10 
 
 9 
 
 3 
 
 3-o 
 
 i-33 
 
 16 
 
 3 
 
 17 
 
 14 
 
 7 
 
 9-5 
 
 3-25 
 
 i? 
 
 3 
 
 14 
 
 ii 
 
 7 
 
 8.0 
 
 2.58 
 
 18 
 
 5 
 
 IS 
 
 IO 
 
 6 
 
 9-5 
 
 2.46 
 
 19 
 
 2 
 
 19 
 
 17 
 
 8 
 
 9.0 
 
 3-25 
 
 20 
 
 2 
 
 16 
 
 14 
 
 7 
 
 6.0 
 
 3-oo 
 
 21 
 
 6 
 
 19 
 
 13 
 
 6 
 
 II.O 
 
 2.38 
 
 22 
 
 8 
 
 .21 
 
 13 
 
 7 
 
 15.0 
 
 2.46 
 
 23 
 
 14 
 
 23 
 
 9 
 
 4 
 
 I9-S 
 
 2.00 
 
 24 
 
 17 
 
 25 
 
 8 
 
 4 
 
 22.0 
 
 1.38 
 
 25 
 
 19 
 
 25 
 
 6 
 
 3 
 
 23-5 
 
 1.29 
 
36 HARVARD-NEWTON BULLETIN 
 
 SELECTING THE SCALE 
 (a) THE BASIS 
 
 In the preceding pages, the following statistical information 
 has been worked out as a basis for selecting the six compositions 
 for the proposed scale : 
 
 A. From the grades given each composition by the twenty-four 
 readers. 
 
 (a) The maximum variation in grades. 
 
 (b) The mean or average grade. 
 
 (c) The median grade. 
 
 B. From the rank or relative position assigned each composi- 
 tion by the twenty-four readers. 
 
 (a) The maximum variation of the whole series. 
 
 (b) The extreme variation of 75 % of the most central cases. 
 
 (c) The median rank. 
 
 (d) The average deviation. 
 
 (b) FIXING STANDARD REQUIREMENTS 
 
 Before selecting the scale on the basis of this information it is 
 necessary to answer these questions. What are the standard or 
 ideal requirements for each of the six compositions of the pro- 
 posed scale in terms of the suggested basis of selection ? What 
 ought the maximum variation, the mean, and the median to be, 
 according to the grade basis of selection ? What ought the two 
 maximum variations, the median rank, and the average deviation 
 to be, according to the rank basis of selection ? We shall now 
 proceed to answer these questions, so far as we can. 
 
 MAXIMUM VARIATIONS AND AVERAGE DEVIATIONS 
 
 Large maximum variation, in either the grades or ranks, 
 indicates great diversity of judgment among the readers, and 
 conversely small maximum variation indicates better agreement 
 in judgment among readers. The former has been shown in the 
 case of composition No. 8, and the latter in the case of composi- 
 tion No. 10 (See p. 34). Obviously, although no definite fixed 
 
HARVARD-NEWTON BULLETIN 37 
 
 standard for variations can be established for any of the six 
 compositions of the proposed scale, it will be agreed that, other 
 things being equal, that composition is most suitable for a place 
 in the scale about which there is greatest agreement in judgment 
 both in grade and rank, as shown by a small maximum variation 
 and a small average deviation. 
 
 THE MEAN, OR AVERAGE GRADE 
 
 The standard requirements of the mean or average grade for 
 each of the six compositions of the scale is not difficult to deter- 
 mine. Our object in this whole study is to secure a scale of six 
 compositions which shall be as nearly as possible, typical, 95 %, 
 85%, 75 %, 65%, 55%, and 45% compositions. Since our 
 ultimate object is expressed in these figures, the six compositions 
 selected ought to have a mean or average grade of 95%, 85%, 
 75%, 65%, 55%, and 45% respectively. 
 
 THE MEDIAN GRADE 
 
 What has been said about the mean or average grade applies 
 also to the median grade. The six compositions for the scale 
 should have respectively a median grade of 95 %, 85 %, 75 %, 65 %, 
 55%, and 45%. 
 
 THE MEDIAN RANK 
 
 The standard requirements for the median rank require some 
 explanation. It is important that the six compositions of the 
 scale should be approximately the same distance apart in rank. 
 A difference of 10% divides them in grades: how many points in 
 rank should divide them ? There are twenty-five ranks, to any 
 one of which a composition may be assigned by any reader. In 
 view of the fact that there are to be six compositions in the scale, 
 there must necessarily be five spaces or steps in the series. The 
 problem then, is to divide the total points in rank between the 
 first and last compositions in the series into five equal divisions 
 so that the six compositions selected for the scale will be equidis- 
 tant one from the other. 
 
38 HARVARD-NEWTON BULLETIN 
 
 It is obvious that beginning with the A or 95 % composition, 
 that composition should have a median rank of i.o. Deducting 
 this one from the twenty-five ranks, the total number, we have 
 twenty-four points of rank remaining to be divided into five equal 
 divisions. This gives 4.8 points in rank for each of the spaces 
 between compositions. By adding this amount 4.8, to i.o, the 
 median rank of the A or 95% composition, we have 5.8, the 
 median rank of the B or 85 % composition of the scale. By add- 
 ing 4.8 to each succeeding median rank we find the following 
 median rank for each of the six compositions. 
 
 Median Rank 
 
 A or 95 % composition i.o 
 
 Bor8s% " 5-8 
 
 C or 75 % 10.6 
 
 Dor 65% 15.4 
 
 E or 55 % " 20.2 
 
 F or 45 % 25.0 
 
 Because the ranks in the graphs are given only in whole num- 
 bers or in whole numbers and halves, it is clear that the decimals 
 in the above median rank numbers cannot be exactly obtained. 
 However, the median rank for each of the compositions in the 
 scale should be as near the above standard median ranks as 
 possible. 
 
 SUMMARY 
 
 From the above discussion, we may summarize the standard 
 requirements for each of the six compositions of the proposed 
 scale. As has been already stated, there is no fixed standard 
 requirement in the case of maximum variations, or average 
 deviation, either in grade or in rank. The only principle to fol- 
 low is this : the narrower the maximum variation the better. 
 
 The other items are summarized as follows : 
 
 Grade or Rank of Mean or Median Median 
 
 Composition in Scale Average Grade Rank 
 
 A Grade or 95 % 95 % 95 % i.o 
 
 B Grade or 85% 85% 85% 5.8 
 
 C Grade or 75 % 75 % 75 % 10.6 
 
 D Grade or 65 % 65 % 65 % 15.4 
 
 E Grade or 55 % 55 % 55 % 20.2 
 
 F Grade or 45 % 45 % 45 % 25.0 
 
HARVARD-NEWTON BULLETIN 39 
 
 (c) MAKING THE SELECTION 
 
 The first step in the selection of the six compositions to con- 
 stitute the scale was to classify the twenty-five compositions into 
 six groups corresponding to the six grades or ranks of composi- 
 tions in the proposed scale. Each composition was placed in 
 that group with which it was most closely identified according to 
 the several statistical items in the table. Such a classification 
 results as follows: 
 
 In general Composition Nos. 
 
 A group of compositions (All from 90% up) 10-15. 
 
 C 
 
 (" 
 (" 
 (" 
 
 1-2-4-8-13-1 7~ I 9~ 2 ' 
 70 " 79%) 16-18-21. 
 
 60 " 69%) 5-6-7-12-22. 
 
 5 " 59%) 9~ I 4-23- 
 
 3-11-24-25. 
 
 Having thus classified the twenty-five compositions, the next 
 step was to select the one from each group which was shown by the 
 standard requirements to be most suitable for the scale. Begin- 
 ning with the A or 95 % composition, we shall consider each group 
 in order. In tabular form are given the numbers of the composi- 
 tions which are to be considered for that particular place in the 
 scale, together with all the statistical data pertaining to each 
 composition. Preceding the table in each case are given the stan- 
 dard requirements for the composition as worked out in the 
 preceding pages. Following the table is the discussion leading 
 to the selection of the most satisfactory composition. 
 
 " A " GRADE OR 95 % COMPOSITION 
 
 Standard Requirements: 
 
 Mean or average grade 95 % 
 
 Median grade 95 % 
 
 Median rank.. . i.o 
 
40 HARVARD-NEWTON BULLETIN 
 
 TABLE VII 
 EXPLANATION 
 
 Letters in the following table designate the following topics: 
 (a) The number of the composition; (b) The highest grade given it; (c) The 
 lowest grade given it; (d) The maximum variation in grade; (e) The mean or aver- 
 age grade; (/) The median grade; (g) The highest rank given it; (h) The lowest 
 rank given it; (i) The maximum variation in rank; (j) The maximum variation 
 in rank of 75 % of the central cases; (k) The median rank; and (I) Average de- 
 viation. 
 
 abed e f g h i j k I 
 io l 95 90 5 94.6 95 ist 3d 2 o i.o .25 
 15 95 80 15 89.6 90 ist loth 9 3 3.0 1.33 
 
 It is not difficult to select the better of these two compositions. 
 No reader felt that composition No. 10 was lower than 90% in 
 grade, or than 3d in rank, whereas No. 15 was graded as low as 
 80%, and ranked as low as loth among the twenty-five composi- 
 tions. The extreme variations are therefore, greater in the 
 case of composition No. 15. Obviously No. 10 was considered by 
 the readers to be more nearly the best composition, hence it is 
 selected as the "A" grade or 95% composition in the scale. 
 No. 15 is too good a composition to be considered for the next 
 composition in the scale, hence it has to be dropped from further 
 consideration. 
 
 " B " GRADE OR 85 % COMPOSITION 
 
 Standard Requirements: 
 
 Mean or Average Grade 85 % 
 
 Median Grade 85 % 
 
 Median Rank 5.8 
 
 From the table below, one sees at once that, judged by these 
 statistical items, several of these compositions are not really 
 competitors for the second place in the scale. Nos. i and 2 are 
 too poor, i. e., they are graded and ranked too low. No. 4 has in 
 its favor the mean and median grades of 84.3% and 85.5% 
 respectively, but it has a wide extreme variation in both grade 
 and rank. Also, its median rank is only 4.5, whereas our require- 
 
 1 Selected for composition No. i in the scale. 
 
HARVARD-NEWTON BULLETIN 41 
 
 TABLE VIII 
 
 EXPLANATION 
 
 Letters in the following table designate the following topics: 
 (a) The number of the composition; (b) The highest grade given it; (c) The 
 lowest grade give it; (d) The maximum variation in grade; (e) The mean or 
 average grade ; (/) The median grade ; (g) The highest rank given it ; (h) The 
 lowest rank given it ; (i) The maximum variation in rank ; (j ) The maximum varia- 
 tion in rank of 75 % of the central cases ; (k) The median rank, and (/) The average 
 deviation. 
 
 a 
 
 b 
 
 c 
 
 d 
 
 e 
 
 / 
 
 I 
 
 h 
 
 I 
 
 j 
 
 k 
 
 / 
 
 i 
 
 95 
 
 68 
 
 27 
 
 81.9 
 
 83.0 
 
 2 d 
 
 i3th 
 
 ii 
 
 8 
 
 7-5 
 
 3.17 
 
 2 
 
 90 
 
 64 
 
 26 
 
 80.0 
 
 80.0 
 
 5th 
 
 1 5th 
 
 10 
 
 4 
 
 7.0 
 
 2.17 
 
 4 
 
 94 
 
 63 
 
 3i 
 
 84.3 
 
 85.5 
 
 2d 
 
 i6th 
 
 14 
 
 6 
 
 4-5 
 
 2.83 
 
 8 
 
 95 
 
 52 
 
 43 
 
 82.3 
 
 85.0 
 
 ISt 
 
 i6th 
 
 15 
 
 10 
 
 6.0 
 
 4.04 
 
 I3 1 
 
 95 
 
 7i 
 
 24 
 
 83-7 
 
 85.0 
 
 ISt 
 
 1 2th 
 
 ii 
 
 7 
 
 6.0 
 
 2.50 
 
 i? 
 
 93 
 
 63 
 
 30 
 
 81.0 
 
 81.5 
 
 3d 
 
 i4th 
 
 ii 
 
 7 
 
 8.0 
 
 2.58 
 
 iQ 
 
 92 
 
 60 
 
 32 
 
 79.6 
 
 80.0 
 
 2 d 
 
 i9th 
 
 17 
 
 8 
 
 9.0 
 
 3-25 
 
 20 
 
 92 
 
 70 
 
 22 
 
 82.7 
 
 85.0 
 
 2d 
 
 1 6th 
 
 14 
 
 7 
 
 6.0 
 
 3.00 
 
 ments call for a median rank of 5.8. No. 8 is one of those com- 
 positions on which teachers cannot agree. Its extreme variation 
 puts it beyond our further consideration here. Comparatively, 
 No. 13 has not a wide variation in grade or rank, and it has a 
 mean and median grade of 83.7 and 85.0 respectively. It is clear 
 that this composition is a strong competitor for the second place. 
 No. 17 has a wider variation than No. 13, and its mean and 
 medians indicate that it is too poor for an 85% composition. 
 No. 19 is likewise too poor. No. 20 has a narrow extreme varia- 
 tion, its mean is a little low, but its median is perfect. Its ex- 
 treme variation by rank is comparatively wide, while its median 
 rank is satisfactory. Clearly the selection of an 85% composi- 
 tion for this scale must be made between Nos. 13 and 20. We 
 select No. 13 because (a) of its slightly narrower extreme varia- 
 tion by grade, because (b) of its mean of 83.7 which is i % nearer 
 the standard than the mean of No. 20, because (c) its extreme 
 variation in rank is three points narrower than that of No. 20 and 
 because, (d) in other respects it is as good as No. 20. 
 
 1 Selected for Composition No. 2 in the scale. 
 
42 HARVARD-NEWTON BULLETIN 
 
 "C" GRADE OR 75% COMPOSITION 
 
 Standard Requirements: 
 
 Mean or Average Grade 75 % 
 
 Median Grade 75 % 
 
 Median Rank . . 10.6 
 
 TABLE IX 
 
 EXPLANATION 
 
 Letters in the following table designate the following topics: 
 (a) The number of the composition; (b) The highest grade given it; (c) The 
 lowest grade given it; (d) The maximum variation in grade; (e) The mean or 
 average grade; (/) The median grade; (g) The highest rank given it; (h) The 
 lowest rank given it; (i) The maximum variation in rank; (;) The maximum 
 variation in rank of 75 % of the central cases; (k) The median rank, and (/) The 
 average deviation. 
 
 a b c d e f g h i j k I 
 
 16 92 68 24 78.2 78.5 3d i7th 14 7 9.5 3.25 
 
 18 90 60 30 79.9 75.0 2d i9th 17 8 9.5 2.46 
 
 21 * 89 54 35 76.1 77.0 6th igth 13 6 n.o 2.38 
 
 Obviously compositions Nos. 16 and 18 are too good composi- 
 tions for typical 75 % grade compositions. The mean or average 
 grade of the former is 78.2 and that of the latter is 79.9. In rank 
 likewise they are better than No. 21 and too good for our purpose. 
 That leaves only No. 21 to be selected for the C grade or 75% 
 composition. Even though this is the best selection that can be 
 made under the circumstances, one could wish that the extreme 
 variations were less and that the other statistical items were 
 nearer the standard requirements. 
 
 " D " GRADE OR 65 % COMPOSITION 
 
 Standard Requirements: 
 
 Mean or Average Grade 65 % 
 
 Median Grade 65 % 
 
 Median Rank 15.4 
 
 1 Selected for Composition No. 3 in the scale. 
 
HARVARD-NEWTON BULLETIN 43 
 
 TABLE X 
 EXPLANATION 
 
 Letters in the following table designate the following topics: 
 (a) The number of the composition; (b) The highest grade given it; (c) The 
 lowest grade given it; (d) The maximum variation in grade; (e) The mean or 
 average grade; (/) The median grade; (g) The highest rank given it; (h) The 
 lowest rank given it; (*) The maximum variation in rank; (j) The maximum 
 variation in rank of 75 % of the most central cases; (k) The median rank, and (/) 
 The average deviation. 
 
 a 
 
 b 
 
 c 
 
 d 
 
 e 
 
 / 
 
 g 
 
 h 
 
 f 
 
 j 
 
 k 
 
 I 
 
 S 
 
 78 
 
 50 
 
 28 
 
 61.1 
 
 60.0 
 
 i4th 
 
 2ISt 
 
 7 
 
 5 
 
 17.0 
 
 1.46 
 
 6 
 
 88 
 
 50 
 
 38 
 
 69.4 
 
 69-5 
 
 8th 
 
 20th 
 
 12 
 
 6 
 
 14.0 
 
 2.0 
 
 7 
 
 80 
 
 40 
 
 40 
 
 63-5 
 
 65.0 
 
 loth 
 
 24th 
 
 14 
 
 7 
 
 16.5 
 
 2.67 
 
 12 
 
 75 
 
 42 
 
 33 
 
 59-9 
 
 60.0 
 
 i4th 
 
 2^d 
 
 9 
 
 4 
 
 17.0 
 
 I.7I 
 
 22 1 
 
 86 
 
 47 
 
 39 
 
 66.6 
 
 66.5 
 
 8th 
 
 2ISt 
 
 13 
 
 7 
 
 15-0 
 
 2.46 
 
 Clearly Nos. 5 and 12 are too poor for a typical 65% composi- 
 tion. No. 7 has a wide extreme variation in both grade and 
 rank: although its mean and median grades might do, its median 
 rank is too low. Hence No. 22 is selected, even though its mean 
 and median grades are slightly high to meet the standard require- 
 ments. 
 
 " E " GRADE OR 55 % COMPOSITION 
 Standard Requirements: 
 
 Mean or Average Grade 55 % 
 
 Median Grade 55 % 
 
 Median Rank 20.2 
 
 TABLE XI 
 EXPLANATION 
 
 Letters in the following table designate the following topics: 
 (a) The number of the composition; (b) The highest grade given it; (c) The 
 lowest grade given it; (d) The maximum variation in grade; (e) The mean or 
 average grade; (/) The median grade; (g) The highest rank given it; (h) The 
 lowest rank given it; (**) The maximum variation in rank; (j) The maximum 
 variation in rank of 75 % of the most central cases; (k) The median rank, and (/) 
 The average deviation. 
 
 a 
 
 b 
 
 c 
 
 d 
 
 e 
 
 / 
 
 g 
 
 h 
 
 i 
 
 j 
 
 k 
 
 | 
 
 9 
 
 75 
 
 40 
 
 35 
 
 56.1 
 
 58.5 
 
 i4th 
 
 2 3 d 
 
 9 
 
 4 
 
 2O.O 
 
 I.S8 
 
 14 
 
 76 
 
 40 
 
 36 
 
 55-4 
 
 53-5 
 
 9th 
 
 24th 
 
 15 
 
 6 
 
 19-5 
 
 2.75 
 
 23 2 
 
 74 
 
 40 
 
 34 
 
 554 
 
 57-5 
 
 i4th 
 
 2 3 d 
 
 9 
 
 4 
 
 19-5 
 
 2.00 
 
 1 Selected for Composition No. 4 in the scale. 
 
 2 Selected as Composition No. 5 in the scale. 
 
44 HARVARD-NEWTON BULLETIN 
 
 In considering No. 9 it is clear that it is a little too good for a 
 typical 55% composition, as shown in two of the three items of 
 the standard requirements. The item of median rank is perfect. 
 However, it should be pointed out that the median rank of the 
 composition selected to occupy a place immediately preceding 
 this was 15.0 instead of the standard requirement of 15.4. Hence, 
 the median rank of composition 55% in the scale should be 
 slightly under 20.0 rather than over it. Between Nos. 14 and 23 
 the advantage of No. 23 is in the narrower extreme variation 
 by rank and in the slightly higher median grade. Hence, No. 23 
 
 is selected. 
 
 " F " GRADE OR 45 % COMPOSITION 
 
 Standard Requirements: 
 
 Mean or Average Grade 45 % 
 
 Median Grade 45 % 
 
 Median Rank 25.0 
 
 TABLE XII 
 
 EXPLANATION 
 
 Letters in the following table designate the following topics: 
 (a) The number of the composition; (b) The highest grade given it; (c) The 
 lowest grade given it; (d) The maximum variation in grade; (e) The mean or 
 average grade; (/) The median grade; (g) The highest rank given it; (h) The 
 lowest rank given it; () The maximum variation in rank; (j) The maximum 
 variation in rank of 75 % of the most central cases; (k) The median rank, and (/) 
 The average deviation. 
 
 a b c d e f g h i j k I 
 
 3 50 30 20 42.7 41.0 2ist 25th 4 2 25.0 .71 
 
 ii 65 40 25 49.5 49.5 i6th 25th 9 5 22.0 2.08 
 
 24 73 30 42 48.9 48.0 1 7th 25th 8 4 22.0 1.38 
 
 25 * 62 20 42 44.9 45.0 i9th 25th 6 3 23.5 1.29 
 
 The objection to No. 3 is that, although all agree that it is 25th 
 composition in rank, or the poorest in the series, the grades given 
 it indicate that it is too poor. On the other hand, No. 1 1 is too 
 good. The choice then lies between Nos. 24 and 25. It is clear 
 that No. 25 more nearly meets the standard requirements. No. 
 24 is a little too good. It should be pointed out that the median 
 
 1 Selected as Composition No. 6 in the scale. 
 
HARVARD-NEWTON BULLETIN 
 
 45 
 
 rank of composition No. 25 is really more satisfactory than at 
 first appears. The median rank of the 55% composition was 
 19.5, a little higher than the standard requirements: if to this is 
 added 4.8, the difference in points of rank between compositions, 
 we have 24.3, a little less than the standard requirements indicate. 
 Hence, No. 25 is selected for the sixth and last composition in the 
 scale. 
 
 SUMMARY 
 
 The following table shows at a glance how the statistical items 
 pertaining to each composition selected for the scale, 1 compare 
 with the standard requirements established for each composition. 
 
 
 
 
 Mean 
 
 
 Median 
 
 
 Median 
 
 
 No. of 
 
 Standard 
 
 Grade of 
 
 Standard 
 
 Grade of 
 
 Standard 
 
 Rank of 
 
 Compositions of the Scale 
 
 Compo- 
 sition 
 Selected 
 
 Mean or 
 Average 
 Grade 
 
 Compo- 
 sition 
 Selected 
 
 Median 
 Grade 
 
 Compo- 
 sition 
 Selected 
 
 Median 
 Rank 
 
 Compo- 
 sition 
 Selected 
 
 A grade or 95 % 
 
 IO 
 
 95 
 
 94.6 
 
 95 
 
 95-0 
 
 I.O 
 
 I.O 
 
 B " 85 % . . 
 
 13 
 
 8c 
 
 83.^ 
 
 85 
 
 8s.o 
 
 c.8 
 
 60 
 
 C " 75% 
 
 21 
 
 7tr 
 
 76.1 
 
 7c 
 
 77 O 
 
 10 6 
 
 II O 
 
 D " 65% 
 
 22 
 
 65 
 
 66.6 
 
 65 
 
 66.5 
 
 15-4 
 
 15.0 
 
 E 55% 
 
 23 
 
 55 
 
 55-4 
 
 55 
 
 57.5 
 
 20.2 
 
 19.5 
 
 F 45% 
 
 25 
 
 45 
 
 44.9 
 
 45 
 
 45-o 
 
 25-0 
 
 23-5 
 
 The chart on page 46 indicates the distribution of ranks for each 
 of the six compositions constituting the description scale. The 
 median is indicated by the line dividing the distribution. For an 
 explanation of this chart see pages 32 and 33. 
 
 DEFINING THE MERITS AND DEFECTS OF THE COMPOSI- 
 TIONS IN THE SCALES 
 
 After the scales had been selected, the next step was to define 
 the merits and defects of each composition in each scale. This 
 was done by a committee consisting of high school and elementary 
 school teachers with the cooperation of the teachers and principals 
 who had originally read the compositions. 
 
 1 For the Compositions of the scale, see pp. 49-57. 
 
4 6 
 
 HARVARD-NEWTON BULLETIN 
 
 
 No 
 
 . 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Ri 
 
 \N 
 
 K 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Nc 
 
 . c 
 
 
 io 
 
 Tip 
 
 OS 
 
 tio 
 
 n 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 -0 
 
 Tip 
 
 OS 
 
 ticn 
 
 
 s 
 
 :a! 
 
 j 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 C 
 
 on 
 
 (p\ 
 
 Bt( 
 
 
 
 
 
 
 
 24 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 i 
 
 et 
 
 
 
 
 
 
 
 
 23 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 22 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 20 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 19 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 18 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 17 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 16 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 IB 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 13 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 N ( 
 
 . 1 
 
 
 
 
 
 12 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 r< 
 
 0. 
 
 K> 
 
 
 
 
 
 
 
 11 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 10 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 21 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 14 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 6 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 21 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 Ni 
 
 . '< 
 
 
 
 
 
 
 
 
 20 
 
 19 
 
 24 
 
 
 
 
 18 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 r> 
 
 0. 
 
 13 
 
 
 
 
 
 
 
 
 
 17 
 
 11 
 
 8 
 
 13 
 
 23 
 
 
 22 
 
 10 
 
 
 16 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 6 
 
 
 12 
 
 1 
 
 2 
 
 3 
 
 15 
 
 4 
 
 9 
 
 5 
 
 14 
 
 7 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 23 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 20 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 N. 
 
 
 
 
 
 
 
 
 
 
 
 
 13 
 
 
 
 24 
 
 22 
 
 
 19 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 No. 
 
 21 
 
 
 
 
 
 
 
 
 
 
 
 
 
 10 
 
 18 
 
 8 
 
 2 
 
 21 
 
 12 
 
 15 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 17 
 
 9 
 
 16 
 
 7 
 
 1 
 
 6 
 
 11 
 
 14 
 
 4 
 
 5 
 
 
 
 
 3 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 22 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 N< 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 19 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 N 
 
 0. 
 
 22 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 20 
 
 11 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 21 
 
 17 
 
 8 
 
 
 
 
 
 12 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 10 
 
 
 
 
 
 16 
 
 15 
 
 7 
 
 
 
 14 
 
 
 5 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 9 
 
 
 13 
 
 
 
 3 
 
 2 
 
 4 
 
 23 
 
 24 
 
 6 
 
 
 1 
 
 18 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 22 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 18 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 16 
 
 
 
 17 
 
 24 
 
 
 
 
 
 
 
 
 
 
 Nj 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 20 
 
 
 13 
 
 21 
 
 
 14 
 
 23 
 
 
 
 
 
 
 
 No. 
 
 23 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 10 
 
 
 6 
 
 11 
 
 
 5 
 
 15 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 8 
 
 
 9 
 
 3 
 
 19 
 
 2 
 
 4 
 
 7 
 
 1 
 
 12 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 24 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 20 
 
 
 
 
 
 
 
 
 
 N 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 23 
 
 
 17 
 
 18 
 
 
 
 
 
 N 
 
 0. 
 
 25 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 21 
 
 22 
 
 15 
 
 14 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 16 
 
 11 
 
 13 
 
 10 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 6 
 
 7 
 
 12 
 
 5 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 19 
 
 8 
 
 9 
 
 3 
 
 4 
 
 2 
 
 1 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
HARVARD-NEWTON BULLETIN 47 
 
 The six compositions of the description scale formed the basis 
 of a teachers' meeting, consisting of eighth grade teachers, ele- 
 mentary school principals, and high school teachers of English. 
 In preparation for the meeting each person was asked to define, in 
 a succinct paragraph, the merits and defects of one of the six 
 compositions in the scale. This assignment had two results: 
 first, it prepared the teachers to participate in the discussion, and 
 second, it furnished five or six paragraphs from as many different 
 people setting forth the merits and defects of each composition 
 as each reader saw them. In further preparation, two persons 
 were asked to lead the discussion on each composition. With this 
 preliminary preparation, the meeting proved exceptionally inter- 
 esting and profitable to the teachers. It was interesting because 
 it " moved," and it was profitable because it was devoted to the 
 consideration of some practical everyday school matters on which 
 every teacher was glad to obtain suggestions. 
 
 At the close of the meeting the written paragraphs mentioned 
 above, were collected and turned over to the committee which was 
 to prepare the necessary paragraphs. The preparation of these 
 paragraphs was not considered an easy task and the committee 
 did its work conscientiously and with thoroughness. After 
 tentative paragraphs had been prepared by each member of the 
 committee according to agreement the committee held a meeting 
 to consider them. Besides the five members of the committee 
 Superintendent Spaulding, Professor Holmes of Harvard, and 
 the Joseph Lee Fellow attended the meeting. The paragraphs 
 were read and discussed at length. It was agreed that the par- 
 agraph accompanying each composition should state, in general, 
 why each composition was better than the one below it in the scale: 
 why it was poorer than the one above it in the scale, and also its 
 own intrinsic worth. With the suggestions from the discussion 
 of the committee, the committee revised the paragraphs and 
 prepared the final definition of merits and defects as they are 
 found accompanying each composition in the description scale. 
 
 Through the courtesy of Mr. Charles S. Thomas, Head of the 
 Department of English in the Classical High School, Newton, and 
 
48 HARVARD-NEWTON BULLETIN 
 
 Editor of the English Leaflets of the New England Association of 
 Teachers of English, the Description Scale and the accompanying 
 paragraphs were published in the February issue of the Leaflet 
 with the request that members of the association offer suggestions 
 for improving them before final publication. 
 
 The same general procedure was followed in Exposition, Argu- 
 mentation, and Narration, except that the teachers were divided 
 into two groups, and each group considered two of the above 
 types. In the case of these scales, also, the final paragraphs were 
 not prepared in time to be presented in the English Teachers 
 Leaflet to obtain the criticism of the members of the association. 
 
HARVARD-NEWTON BULLETIN 65 
 
 Comparison 
 
 The theme is ranked higher than No. 4 because it is more coherent, and 
 the sentence structure is better. Because the sentence structure is more 
 monotonous, and because it is not so well punctuated, the theme is placed 
 lower than No. 2. 
 
66 HARVARD-NEWTON BULLETIN 
 
 No. 4. "D" GRADE COMPOSITION. VALUE, 61.0% 
 How TO BUILD A FIRE l 
 
 To build a furnace fire you must as the following instructs. 
 First, you must rake down and take out all the cinders, clin- 
 kers and ashes, sift them, so as to do away with clinkers and 
 ashes and use the cinders for some other purpose. 
 5 The next step, is to get your material ready to work with. 
 First you put some paper in the furnace, on that some wood, 
 wait until the wood is blazing, then take three good-sized 
 shovels full of small coal and put it on that. Take the same 
 amount of big coal repeating the same process. A fire like 
 10 this is supposed to last eight hours. The next thing to do is 
 to open all your drafts to let the air in and all the gas and 
 smoke out so as the fire will progress quicker. Last of all 
 investigate your work to see if it is allright. Your task is 
 done. 
 
 Merits 
 
 This theme shows clear thinking and a mastery of the details that make 
 the whole convincing. The writer has evidently built furnace fires and re- 
 lives the process as he writes. His vocabulary is throughly suited to his 
 theme. 
 
 Defects 
 
 The writer has shown no care in revising his theme. There is an obvious 
 omission in the first line, careless sentence structure in the second para- 
 graph, and a misspelling of all right in the last line. The sentence com- 
 mencing A fire like this is an intrusion an offense against coherence. 
 The last sentence may be omitted. 
 
 Comparison 
 
 This theme is thinner in subject-matter than No. 3 and is more obviously 
 lacking in a sense of style and form. A firm grasp of the elementary prin- 
 ciples of composition secures its place above No. 5. 
 
 1 Composition No. 6 in the complete set. 
 
HARVARD-NEWTON BULLETIN 67 
 
 No. 5. " E " GRADE COMPOSITION. VALUE, 52.6% 
 How TO HARNESS A HORSE 1 
 
 When you get the horse out of his stall, the first thing to do 
 is to get the bridal on him. On putting the bridal on you 
 first take hold of the piece of the bridal that goes around his 
 neck and then take hold of the bit in the other hand. You 
 then take hold of the lower part of his mouth and pull it 5 
 down and then slip the bit in his mouth. The bit is the iron 
 piece on the bridal. When you get the bit in his mouth you 
 tie the strap around his neck. Then take the collar and put 
 it over his head on to his neck. The last thing to do is to put 
 the saddle on, throw the saddle over back and buckle the 10 
 inside strap on the to the other side of the saddle, you put it 
 under his stomack to buckle it and you have to have it good 
 and tight. 
 
 Merits 
 
 The merits are three in number: (i) the writer knew how to harness a 
 horse; (2) he has given his directions in an orderly manner; (3) he has 
 kept to his subject has omitted all unnecessary details. 
 
 Defects 
 
 The defects are as follows: (i) the ideas are crudely and awkwardly 
 expressed; (2) the vocabulary is meagre, as shown for example, by the fre- 
 quent repetition of bridle and saddle; (3) bridle and stomach are misspelled; 
 and (4) the last sentence is ungrammatical and crude it should be recast. 
 
 Comparison 
 
 This theme precedes No. 6 in the scale because it is superior in sentence 
 structure, spelling, and clearness. It is inferior to No. 4 in vocabulary and 
 mechanics. 
 
 1 Composition No. 8 in the complete set. 
 
68 HARVARD-NEWTON BULLETIN 
 
 No. 6. "F" GRADE COMPOSITION. VALUE, 39.1% 
 How TO HARNESS A HORSE l 
 
 I am going to tell you how to harness a horse. First you 
 led him out of his stall take off his halter then put on his col- 
 lar then put on his bridle put on his saddle. Then led him out 
 to the team back him up into the shafts left them up put 
 5 them into the socets on the saddle them tient up the shaft 
 strap then fix the traces to the whippel tree the the bretchen 
 then the chect rains the run the rains throught the holds in 
 the saddle back to the team. 
 
 Merits 
 
 Only two things can be said in commendation of this theme: the first 
 sentence is free from mechanical errors, and the directions for harnessing a 
 horse are given in logical order. 
 
 Defects 
 
 The theme has many faults. The meagerness of the writer's vocabulary, 
 made conspicuous by the childish repetition of then and put; his utter lack 
 of sentence sense; his inexcusable errors in spelling, such as led for lead, 
 the for then, left for lift, tient for tighten all these illiteracies make the 
 theme rank low. 
 
 Comparison 
 
 This theme is inferior to No. 5 in spelling, punctuation, and sentence 
 structure. With its bare enumeration of the steps in the process of harnessing 
 a horse, it completely lacks the slight elements of grace which dependent 
 clauses and fuller explanations contribute to No. 5. 
 
 1 Composition No. u in the complete set. 
 
HARVARD-NEWTON BULLETIN 69 
 
 THE ARGUMENTATION SCALE 
 
 THE TABULATION OF RESULTS OF READINGS l 
 
 Table I shows the grade that each of the twenty-four readers 
 gave each of the twenty-five compositions in argumentation. 
 
 TABLE I 
 
 Number of Composition 
 6 8 
 ZP4 12 3 4 5 6 7 8 Q 10 ii 12 13 14 15 16 17 18 ig 20 21 22 23 24 25 
 
 52 73 48 78 78 70 88 35 oo 60 82 48 58 55 80 95 58 40 oo 55 85 45 48 50 78 
 
 66 69 60 72 77 7i 88 67 65 75 83 78 57 82 87 97 93 45 92 68 79 50 40 74 86 
 
 50 50 40 75 75 70 85 45 60 80 70 50 50 55 65 95 65 45 95 60 55 45 40 55 oo 
 
 45 50 .. * .... 75 60 70 70 60 45 40 80 85 40 .. 75 70 80 .. 40 55 85 
 
 75 70 58 80 .. . . 65 45 85 68 83 55 50 88 . . 90 72 50 78 68 65 58 62 60 78 
 
 70 62 50 75 83 70 88 55 92 80 92 57 58 72 73 85 65 66 95 55 80 60 52 69 94 
 
 65 70 60 85 80 80 75 68 85 70 80 75 70 65 70 95 75 60 85 70 68 65 55 65 70 
 
 68 67 52 83 90 85 73 48 78 50 80 49 46 63 60 92 55 64 95 70 87 65 45 57 75 
 
 60 60 45 50 65 70 65 50 70 60 60 50 40 60 75 95 65 45 85 60 50 45 45 70 65 
 
 70 70 70 87 90 85 93 60 93 80 94 60 50 82 80 95 65 73 90 85 89 78 65 68 93 
 
 55 85 45 80 70 70 90 45 95 65 80 70 50 55 . . 95 75 55 90 75 65 50 45 65 75 
 
 70 50 55 52 62 69 80 66 85 73 75 74 40 45 61 90 51 54 83 82 87 63 48 57 88 
 
 55 60 45 65 70 80 70 45 75 60 80 80 45 50 85 90 50 50 85 80 85 65 45 60 90 
 60 85 45 55 65 48 75 45 oo 65 75 79 55 55 90 95 45 50 90 63 70 50 45 55 85 
 
 56 54 45 85 80 83 91 78 88 65 70 58 50 68 78 95 54 46 93 52 72 50 46 50 75 
 
 60 55 45 58 . . 50 60 50 95 50 60 65 55 65 . . 95 55 50 88 70 70 50 50 55 85 
 
 61 56 54 82 69 78 79 55 94 75 93 73 65 80 90 95 67 60 88 70 85 58 48 63 92 
 
 51 74 45 70 58 67 87 45 95 54 89 71 49 56 70 95 52 60 82 65 77 47 50 68 85 
 48 48 40 50 50 65 68 55 50 52 56 50 45 48 65 oo 68 55 85 52 48 48 42 50 70 
 
 67 69 60 72 70 7i 85 66 65 76 84 78 58 82 87 95 73 50 oo 68 80 55 45 74 88 
 65 63 56 78 74 72 85 57 89 67 75 62 60 74 73 95 64 58 90 62 82 45 40 65 76 
 61 71 45 87 85 77 93 52 90 65 75 59 55 64 76 95 66 48 92 54 78 53 46 47 79 
 58 52 43 88 83 80 91 47 85 66 70 55 53 63 75 95 61 45 93 So 78 49 4O 46 7* 
 
 1 Mr. William J. Sands did the statistical work on the Argumentation Scale. 
 
 2 Reader No. 4 did not read this set of compositions. 
 
 3 Blank spaces indicate that the reader did not grade that composition because 
 the composition was not considered suitable for a scale. 
 
70 HARVARD-NEWTON BULLETIN 
 
 Table IV shows the ranking of each composition in argumenta- 
 tion by each reader. 
 
 TABLE IV 
 
 "o-o Choice in Rank 
 
 p* 
 
 
 
 
 
 
 6 
 
 7 
 
 8 
 
 Q 
 
 IO 
 
 // 
 
 12 
 
 13 
 
 14 
 
 15 
 
 16 
 
 77 
 
 18 
 
 19 
 
 20 
 
 21 
 
 22 
 
 33 
 
 34 
 
 35 
 
 I 
 
 16 
 
 9 
 
 19 
 
 7 
 
 21 
 
 ii 
 
 IS 
 
 25 
 
 5 
 
 4 
 
 2 
 
 6 
 
 10 
 
 13 
 
 17 
 
 20 
 
 14 
 
 i 
 
 24 
 
 23 
 
 12 
 
 3 
 
 22 
 
 18 
 
 8 
 
 2 
 
 16 
 
 19 
 
 7 
 
 IS 
 
 25 
 
 ii 
 
 14 
 
 21 
 
 12 
 
 S 
 
 10 
 
 24 
 
 17 
 
 4 
 
 6 
 
 2 
 
 2O 
 
 8 
 
 I 
 
 9 
 
 3 
 
 13 
 
 22 
 
 18 
 
 23 
 
 ?1 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 O 
 
 4! 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 5 
 
 16 
 
 25 
 
 21 
 
 IS 
 
 19 
 
 7 
 
 ii 
 
 10 
 
 20 
 
 9 
 
 12 
 
 24 
 
 2 
 
 I 
 
 13 
 
 17 
 
 14 
 
 23 
 
 18 
 
 8 
 
 6 
 
 5 
 
 3 
 
 22 
 
 4 
 
 6 
 
 16 
 
 14 
 
 9 
 
 ii 
 
 4 
 
 25 
 
 19 
 
 I 
 
 I? 
 
 2 
 
 10 
 
 20 
 
 7 
 
 21 
 
 23 
 
 24 
 
 3 
 
 22 
 
 12 
 
 13 
 
 18 
 
 8 
 
 S 
 
 6 
 
 IS 
 
 7 
 
 19 
 
 25 
 
 ii 
 
 9 
 
 7 
 
 16 
 
 5 
 
 10 
 
 21 
 
 4 
 
 IS 
 
 14 
 
 6 
 
 I 
 
 24 
 
 18 
 
 17 
 
 2 
 
 22 
 
 13 
 
 12 
 
 8 
 
 20 
 
 23 
 
 3 
 
 8 
 
 16 
 
 19 
 
 9 
 
 4 
 
 ii 
 
 6 
 
 7 
 
 12 
 
 17 
 
 20 
 
 10 
 
 25 
 
 S 
 
 IS 
 
 13 
 
 2 
 
 21 
 
 8 
 
 24 
 
 22 
 
 I 
 
 14 
 
 18 
 
 3 
 
 23 
 
 9 
 
 19 
 
 16 
 
 5 
 
 21 
 
 6 
 
 4 
 
 ii 
 
 9 
 
 25 
 
 7 
 
 20 
 
 I 
 
 2 
 
 22 
 
 18 
 
 14 
 
 IS 
 
 24 
 
 17 
 
 3 
 
 10 
 
 12 
 
 8 
 
 13 
 
 23 
 
 IO 
 
 16 
 
 19 
 
 IS 
 
 24 
 
 9 
 
 6 
 
 25 
 
 17 
 
 7 
 
 5 
 
 20 
 
 14 
 
 II 
 
 10 
 
 I 
 
 2 
 
 21 
 
 12 
 
 8 
 
 4 
 
 13 
 
 22 
 
 18 
 
 3 
 
 13 
 
 ii 
 
 16 
 
 ii 
 
 7 
 
 9 
 
 25 
 
 S 
 
 19 
 
 21 
 
 4 
 
 6 
 
 20 
 
 14 
 
 10 
 
 25 
 
 22 
 
 18 
 
 I 
 
 2 
 
 3 
 
 24 
 
 17 
 
 23 
 
 8 
 
 12 
 
 13 
 
 12 
 
 9 
 
 16 
 
 7 
 
 19 
 
 2 
 
 ii 
 
 4 
 
 20 
 
 17 
 
 25 
 
 12 
 
 5 
 
 6 
 
 IO 
 
 24 
 
 21 
 
 18 
 
 14 
 
 I 
 
 13 
 
 22 
 
 8 
 
 3 
 
 23 
 
 IS 
 
 13 
 
 16 
 
 25 
 
 21 
 
 9 
 
 19 
 
 20 
 
 7 
 
 II 
 
 12 
 
 IO 
 
 J 
 
 6 
 
 8 
 
 22 
 
 5 
 
 IS 
 
 24 
 
 3 
 
 18 
 
 4 
 
 17 
 
 2 
 
 23 
 
 14 
 
 13 
 
 14 
 
 16 
 
 25 
 
 21 
 
 IS 
 
 19 
 
 2O 
 
 ii 
 
 12 
 
 6 
 
 9 
 
 S 
 
 7 
 
 4 
 
 22 
 
 2 
 
 10 
 
 24 
 
 I 
 
 14 
 
 17 
 
 18 
 
 3 
 
 8 
 
 13 
 
 23 
 
 IS 
 
 16 
 
 9 
 
 IS 
 
 19 
 
 2 
 
 25 
 
 12 
 
 7 
 
 ii 
 
 21 
 
 s 
 
 IO 
 
 20 
 
 I 
 
 14 
 
 24 
 
 4 
 
 13 
 
 18 
 
 22 
 
 6 
 
 8 
 
 17 
 
 23 
 
 3 
 
 16 
 
 16 
 
 19 
 
 7 
 
 9 
 
 4 
 
 6 
 
 5 
 
 15 
 
 25 
 
 21 
 
 II 
 
 14 
 
 10 
 
 17 
 
 12 
 
 I 
 
 2 
 
 20 
 
 13 
 
 24 
 
 22 
 
 8 
 
 18 
 
 23 
 
 3 
 
 I7 1 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 18 
 
 16 
 
 9 
 
 ii 
 
 25 
 
 5 
 
 19 
 
 21 
 
 4 
 
 14 
 
 7 
 
 6 
 
 10 
 
 12 
 
 20 
 
 5 
 
 17 
 
 13 
 
 24 
 
 I 
 
 18 
 
 22 
 
 2 
 
 8 
 
 3 
 
 23 
 
 19 
 
 16 
 
 9 
 
 ii 
 
 7 
 
 25 
 
 19 
 
 21 
 
 2 
 
 12 
 
 4 
 
 IS 
 
 24 
 
 6 
 
 20 
 
 18 
 
 S 
 
 14 
 
 10 
 
 17 
 
 i 
 
 23 
 
 13 
 
 8 
 
 22 
 
 3 
 
 10 
 
 16 
 
 19 
 
 25 
 
 7 
 
 i? 
 
 6 
 
 IS 
 
 II 
 
 8 
 
 18 
 
 20 
 
 10 
 
 4 
 
 5 
 
 9 
 
 12 
 
 24 
 
 14 
 
 21 
 
 22 
 
 I 
 
 2 
 
 IS 
 
 23 
 
 3 
 
 21 
 
 16 
 
 19 
 
 25 
 
 IS 
 
 7 
 
 ii 
 
 14 
 
 21 
 
 12 
 
 10 
 
 24 
 
 17 
 
 4 
 
 6 
 
 5 
 
 2 
 
 20 
 
 I 
 
 8 
 
 9 
 
 3 
 
 13 
 
 22 
 
 18 
 
 23 
 
 22 
 
 16 
 
 19 
 
 9 
 
 7 
 
 21 
 
 4 
 
 25 
 
 II 
 
 5 
 
 14 
 
 IS 
 
 6 
 
 10 
 
 24 
 
 i 
 
 17 
 
 2 
 
 12 
 
 20 
 
 13 
 
 18 
 
 8 
 
 3 
 
 22 
 
 23 
 
 23 
 
 16 
 
 7 
 
 19 
 
 9 
 
 4 
 
 S 
 
 25 
 
 21 
 
 6 
 
 IS 
 
 II 
 
 2 
 
 17 
 
 IO 
 
 14 
 
 I 
 
 12 
 
 13 
 
 20 
 
 22 
 
 8 
 
 18 
 
 24 
 
 23 
 
 3 
 
 24 
 
 16 
 
 19 
 
 7 
 
 4 
 
 9 
 
 5 
 
 6 
 
 21 
 
 IS 
 
 25 
 
 II 
 
 10 
 
 14 
 
 17 
 
 i 
 
 12 
 
 13 
 
 2 
 
 2O 
 
 22 
 
 8 
 
 24 
 
 18 
 
 3 
 
 23 
 
 1 Reader No. 4 did not read these compositions. Readers No. 3 and 17 gave 
 several compositions the same grade, hence the compositions could not be ranked. 
 
HARVARD-NEWTON BULLETIN 71 
 
 THE COMPLETED ARGUMENTATION SCALE 
 No. i. "A" GRADE COMPOSITION. VALUE, 93.2% 
 
 Is A YEAR OF TRAVEL MORE ADVANTAGEOUS THAN A 
 YEAR OF STUDY ? l 
 
 Yes. In nearly all lines of study, travel is an important 
 aid. 
 
 Geography for instance. Places dryly discussed in class 
 and soon forgotten become actual. Races that made no im- 
 pression in print are fastened in the mind by a hundred vivid 5 
 pictures. Products that seemed so useless and annoying to 
 learn are full of importance and interest when seen being 
 loaded on ships or mules. Queer oriental customs, laughed 
 at when studied become natural, comprehended facts. Sta- 
 tistics are seldom learned while travelling, but when they are, 10 
 they are remembered because the bare, dry fact is connected 
 with an actual thing not merely figures, groaned over be- 
 cause they were difficult to learn. 
 
 History also. Some spot where General A's army ma- 
 noeuvered in this direction, and Colonel B's in that, becomes 15 
 real ground, to be reverenced and remembered. The hills 
 that would get on the wrong side of the diagram are here in 
 their true position, and that granite monument marks the 
 spot where the brave commanders died. 
 
 Languages, too. An impossible French name slides off 20 
 glibly when often heard. Some word is gained with very lit- 
 tle effort, another, and the door to the whole language is 
 thrown wide. The same is true of Italian, and the best part 
 of it is that the pronunciation is correct. 
 
 Mathematics could not be so easily acquired, it is true, as 25 
 in the schoolroom. But when the mind has been broadened 
 by the year of travel, the student would more easily learn this 
 branch. 
 
 1 Composition No. 1 6 in the complete set. 
 
72 HARVARD-NEWTON BULLETIN 
 
 As for English, a year like this would be the greatest ad- 
 30 vantage possible. The lovely landscapes of Switzerland or 
 Italy, the odd manners of the Orientals, the thousand types 
 of humanity everything lends its aid. 
 
 On the whole, travel is a thousand times an improvement 
 over the school, because it creates interest, stimulates 
 35 thought, and broadens the traveller as nothing else can. 
 
 Merits 
 
 This theme is excellent because the writer has his argument clearly in 
 mind, follows a carefully planned outline, and expresses himself convincingly. 
 The question is stated clearly, and the conclusion leaves a definite impres- 
 sion on the reader's mind. The illustrations used are well chosen and 
 vivid, especially the ones about foreign customs and historical scenes. 
 Vocabulary and sentence form are good. 
 
 Defects 
 
 A different arrangement of arguments in the order of climax would add 
 strength to the theme. The beginning of each paragraph is awkward 
 because of the use of detached phrases and words: i.e. yes; geography for 
 instance, languages too. Either these phrases should be deliberately iso- 
 lated as headings, and the whole paper thrown into outline form, or else 
 they should be expanded into complete statements. There is some weakness 
 in punctuation, as for example, a semicolon should be substituted for the 
 comma after effort in line 22. 
 
 Comparison 
 
 This theme is better than No. 2 because it follows a more definite plan, 
 its arguments are more clearly stated, and its vocabulary is more varied. 
 
HARVARD-NEWTON BULLETIN 73 
 
 No. 2. "B" GRADE COMPOSITION. VALUE, 81.5% 
 WHY COUNTRY LITE is BETTER THAN CITY LIFE l 
 
 Country life is the best kind of life there is to be lived. 
 
 The country is good for children and grown up people alike. 
 The children have plenty of fresh air and room to run about 
 in. To be sure they do not have the large schools of the city; 
 but they have the great school of Nature and that is greater 5 
 than any learning they could get in city schools. The parents 
 are not afraid to let them run about because there are no 
 automobiles, electric cars, etc., to run over them. There is 
 safety in the country. 
 
 The country is the best place sick people could go. The 10 
 air is pure and bracing and there is water cool and fresh from 
 the spring or well. For the tired man working in a busy 
 city office a rest in the country is the best thing for him and 
 the sooner he gets it the better it is for him. 
 
 The country people have good times even if they don't go 15 
 to fancy balls, theatres, etc., and stay up until one or two 
 o'clock in the morning. They have a good healthy time and 
 are home and in bed by nine or half -past. 
 
 For these reasons, and many others, the country is the best 
 place for children or adults, sick or well. 20 
 
 Merits 
 
 The writer has a sense of arrangement, shown (i) by his introduction; 
 (2) by his paragraphs that develop successively the advantages of country 
 life; and (3) by his conclusion. The composition is also correct in its 
 elementary mechanics spelling, punctuation, grammar, and paragraph 
 and sentence structure. The vocabulary, while not distinctive, is adequate. 
 
 Defects 
 
 The arguments are not wholly convincing. For example, the great school 
 of Nature is, for most children, not a satisfactory substitute for a city school. 
 Then, to say without qualification, that the country is the best place for 
 
 1 Composition No. 25 in the complete set. 
 
74 HARVARD-NEWTON BULLETIN 
 
 sick people to go, is misleading. The repeated use of etc. is ineffective. 
 The summary should not contain the phrase and many others. 
 
 Comparison 
 
 This composition lacks the virility of thought and diction that marks 
 No. i. It is, however, superior to No. 3 in arrangement, in sentence struc- 
 ture, and in its firmness of mental grasp. 
 
HARVARD-NEWTON BULLETIN 75 
 
 No. 3. " C " GRADE COMPOSITION. VALUE, 71.9% 
 THE GOOD OF A LANTERN IN GEOGRAPHY 1 
 
 A lantern is a good thing in a school as it shows the differ- 
 ent cities, countries, and also the ways of the people. This 
 gets the children acquainted with foreign places. Also the 
 reciting about the pictures gives them an interest which 
 would not be so great as if they were in the schoolroom. 5 
 
 It is a good thing for the children to get up before the class 
 and recite, that gives them self confidence and also prepares 
 them more for high school. In having the lantern it interests 
 the children so they look for pictures to use, and they find and 
 read many interesting things. 10 
 
 When the pictures are passed out they look up many 
 things at home and find out new things and so as other chil- 
 dren find things out they help the class. 
 
 Merits 
 
 The merits of this composition are three: (i) the form is purely argumen- 
 tative throughout; (2) the pupil had a few distinct advantages of the use 
 of a lantern in mind; and (3) troublesome words like acquaintance and 
 foreign are correctly spelled. 
 
 Defects 
 
 The defects are as follows: (i) poor arrangement of facts makes this 
 theme ineffective; (2) the sentences are awkward for example, the third 
 and fifth; (3) the fourth sentence illustrates the " run-on " sentence or 
 " comma error "; (4) there is lack of clearness in the antecedents of pro- 
 nouns they in lines 5, n, and 13; (5) there is a poverty of vocabulary 
 the word thing or things occurs six times, and children occurs in almost 
 every sentence. 
 
 Comparison 
 
 This theme follows No. 2 in the scale because it has less mature thought, 
 inferior sentence structure, and a more limited and childish vocabulary. 
 It precedes No. 4 because, although not so effectively written, it adheres more 
 closely to the form of writing required argumentation. 
 
 1 Composition No. 6 in the complete set. 
 
76 HARVARD-NEWTON BULLETIN 
 
 No. 4. "D" GRADE COMPOSITION. VALUE, 63.6% 
 FOOTBALL is A BETTER GAME THAN BASEBALL l 
 
 Football is a better game than baseball. It's not only 
 
 (only) muscle and strength that makes a football player. He 
 
 must have the power to think quick. When the team is a 
 
 short distance from the goal and scores, somebody's quick 
 
 5 thinking has done it. 
 
 Baseball is an exciting game in parts. But football is ex- 
 citing from the (the) time the ball is kicked till the referee 
 blows his whistle. Football shows when a man or boy has 
 grit. I once attended a game. The score was tie and the 
 
 10 opposing team were only a few yards from the goal and three 
 minutes to play. The quarterback gave his signal and they 
 struck the line. The struggling mass moved a few feet and 
 stopped with a jump. Out of the mass a man jumped. His 
 face was all battered, but he struggled on till he had got the 
 
 15 ball over the line. Some people call it brutal. Some cases it 
 is and some it is not. I would rather strike the line in foot- 
 ball clothes for a few yards gain, than stand up at a plate and 
 have a man hurl baseballs at me. Which would you rather 
 do? 
 
 Merits 
 
 The writer of this theme evidently has a plan of procedure in mind. 
 After clearly stating his proposition in the first sentence, he seeks to establish 
 three good points of superiority for the game of football. Though unable 
 to prove these points, he has succeeded in leaving with the reader a firm 
 conviction of his belief that football is a good sport. Lines 9-15 give effec- 
 tively the writer's impressions of a critical moment in a game. 
 
 Defects 
 
 The author fails to prove his proposition, because he mistakes for argu- 
 ments a few loosely stated facts and preferences. His plan of reasoning, 
 though suggested, is entirely undeveloped. Mechanical errors are numer- 
 ous: (i) two mistakes in the use of parentheses (lines 2 and 7); (2) plural 
 verb with singular subject (line 10) ; (3) adjective modifies a verb (line 3) ; 
 
 1 Composition No. a in the complete set. 
 
HARVARD-NEWTON BULLETIN 77 
 
 and the pronoun it is used vaguely in line 15. The paragraphs are 
 without system, and the sentences in general are short and disconnected. 
 
 Comparison 
 
 This composition deserves its place in the scale above No. 5 because of 
 its superior mechanical accuracy and larger vocabulary. It does not hold 
 as closely to its subject as does No. 3. 
 
78 HARVARD-NEWTON BULLETIN 
 
 No. 5. "E" GRADE COMPOSITION. VALUE, 52.3% 
 WHY EVERY GIRL SHOULD LEARN TO COOK 1 
 
 Every girl should learn to cook. If she was out of imploi- 
 ment it would be very useful. 
 
 Later on, if she should keep house she would want to be 
 able to cook. If she didn't know how, the lady would have 
 5 to hire a cook. Sometime, perhaps, she could not pay to 
 keep a cook. Then she would say, " Oh, how I wished I 
 learned to cook when I was younger, instead of buying bak- 
 ers things." 
 
 If this womans children wanted to learn to cook who 
 10 would they ask to learn them. Their mother couldn't, be- 
 cause she did n't know how her self. So you may see for 
 yourself it is very important that every girl should learn to 
 cook. 
 
 Merits 
 
 This theme shows that the author had a plan; for it gives several sound 
 reasons why every girl should learn to cook reasons arranged in a fairly 
 logical order. The punctuation of certain sentences is unexpectedly good. 
 
 Defects 
 
 The theme is, however, faulty in unity and coherence. There is a shift- 
 ing of tense, a lack of definite antecedents for such words as it (line 2), and 
 lady (line 4). The arguments are not sufficiently expanded to make them 
 clear. The theme is further marred by such mechanical errors as : misspelling 
 of employment (line i) ; a complete absence of the apostrophe and interroga- 
 tion point; the incorrect use of who (line 9), and learn (line 10). 
 
 Comparison 
 
 This theme is placed lower in the scale than No. 4; it has less force as an 
 argument, and has more mechanical errors. 
 
 1 Composition No. 13 in the complete set. 
 
HARVARD-NEWTON BULLETIN 79 
 
 No. 6. " F " GRADE COMPOSITION. VALUE, 47.0% 
 THE WINTER is THE BEST SEASON l 
 
 The winter is the best season in the year because they are 
 skating and pung-riding coasting etc. In the spring when 
 you plant the crops you are kept busy. In the summer tak- 
 ing care of them weeding in them and picking bean peas, 
 and the only day you get of is Sunday. 5 
 
 In July it is too hot you swet all the time the horse won't 
 run because it is so hot. In the fall you are busy get in all the 
 crops and cutting corn banking celery, digging potatoes, pull- 
 ing turnips, picking apples, pears, cutting onions and every 
 thing else. So the winter is the only time you can have fun 10 
 when they are nothing to do. 
 
 Merits 
 
 There is little to defend in this theme. Only one line of argument is 
 attempted; that one, however, is arranged logically according to the order 
 of the four seasons. The use of specific words is good, and the large number 
 of illustrations drawn from the pupil's own experience arouses sympathy in 
 the reader. 
 
 Defects 
 
 Although the range of arguments is limited, and those arguments consist 
 mainly of the enumeration of the woes that beset a boy in the other seasons 
 rather than in the joys of winter, it is from the mechanical standpoint that 
 this theme is a failure. There is manifested an entire lack of sentence-sense 
 since examples of both " run-on " sentences and unfinished clauses occur 
 frequently. Errors in spelling, punctuation, and grammar are found in 
 every line. The whole composition is the expression of a backward, limited 
 mind. 
 
 Comparison 
 
 The theme deserves a lower place in the scale than No. 5 because of the 
 greater frequency of mechanical errors. In weight of subject matter, how- 
 ever, it is superior. 
 
 1 Composition No. 23 in the complete set. 
 
8o 
 
 HARVARD-NEWTON BULLETIN 
 
 THE NARRATION SCALE 
 
 THE TABULATION or RESULTS OF READINGS l 
 
 Table I shows the grade that each of the twenty-four readers 
 gave each of the twenty-five compositions in narration. 
 
 TABLE I 
 
 ""! 
 
 6 S 
 
 
 
 
 
 Number of the Composition 
 S 6 7 8 9 10 ii 12 13 14 15 16 17 18 
 
 IQ 
 
 20 
 
 21 
 
 22 
 
 23 
 
 24 25 
 
 i 
 
 2 
 3 
 
 73 
 90 
 93 
 
 45 
 58 
 45 
 
 95 
 95 
 95 
 
 68 
 58 
 
 68 58 9 58 55 70 60 80 73 68 75 55 50 78 
 61 55 56 58 59 60 76 87 69 68 73 62 72 75 
 58 65 93 45 60 60 77 90 50 55 65 68 80 83 
 
 85 
 82 
 73 
 
 60 
 6 4 
 
 4 8 
 
 75 
 86 
 75 
 
 75 
 80 
 90 
 
 65 
 65 
 68 
 
 73 80 
 70 85 
 88 85 
 
 5 
 
 75 
 
 50 
 
 95 
 
 78 
 
 73 55 70 50 48 70 73 85 88 68 75 85 70 65 
 
 90 
 
 60 
 
 90 
 
 85 
 
 80 
 
 80 93 
 
 6 
 
 79 
 
 46 
 
 93 
 
 66 
 
 83 64 75 60 78 52 58 95 78 82 80 85 73 70 
 
 75 
 
 67 
 
 60 
 
 88 
 
 75 
 
 77 91 
 
 7 
 
 78 
 
 48 
 
 95 
 
 65 
 
 75 55 9 2 67 67 60 62 87 70 68 71 60 64 82 
 
 9i 
 
 58 
 
 88 
 
 72 
 
 66 
 
 68 90 
 
 8 
 
 70 
 
 50 
 
 80 
 
 65 
 
 68 50 75 . . 168 60 50 65 70 60 50 50 50 65 
 
 65 
 
 60 
 
 68 
 
 65 
 
 60 
 
 50 70 
 
 9 
 
 70 
 
 45 
 
 95 
 
 75 
 
 55 53 85 73 72 72 73 90 83 57 48 63 78 80 
 
 87 
 
 50 
 
 73 
 
 68 
 
 65 
 
 60 93 
 
 10 
 
 75 
 
 45 
 
 95 
 
 73 
 
 65 55 85 70 68 63 70 90 80 58 45 67 70 80 
 
 80 
 
 4 8 
 
 70 
 
 68 
 
 73 
 
 58 90 
 
 ii 
 
 78 
 
 40 
 
 90 
 
 65 
 
 87 63 95 70 75 60 73 84 73 75 55 55 73 80 
 
 83 
 
 60 
 
 86 
 
 76 
 
 70 
 
 78 85 
 
 12 
 
 90 
 
 45 
 
 95 
 
 83 
 
 67 48 70 57 53 55 80 93 92 63 75 65 72 85 
 
 75 
 
 50 
 
 75 
 
 70 
 
 78 
 
 60 88 
 
 13 
 
 82 
 
 65 
 
 94 
 
 81 
 
 79 78 93 80 77 88 89 87 91 76 90 60 74 75 
 
 92 
 
 70 
 
 86 
 
 83 
 
 84 
 
 85 95 
 
 14 
 
 88 
 
 45 
 
 95 
 
 70 
 
 85 65 90 55 50 65 75 90 85 70 90 75 78 80 
 
 95 
 
 75 
 
 85 
 
 80 
 
 85 
 
 78 95 
 
 15 
 
 70 
 
 45 
 
 95 
 
 50 
 
 64 54 91 65 75 58 62 83 68 60 75 50 68 78 
 
 85 
 
 49 
 
 85 
 
 68 
 
 59 
 
 55 88 
 
 16 
 
 78 
 
 45 
 
 92 
 
 56 
 
 65 50 90 62 58 60 60 88 72 66 68 57 62 80 
 
 85 
 
 52 
 
 87 
 
 75 
 
 68 
 
 70 95 
 
 i? 
 
 92 
 
 45 
 
 93 
 
 48 
 
 83 50 70 55 60 68 80 83 75 78 80 50 58 94 
 
 93 
 
 68 
 
 85 
 
 88 
 
 58 
 
 85 85 
 
 18 
 
 88 
 
 45 
 
 95 
 
 68 
 
 71 55 88 60 70 62 76 92 83 73 75 53 65 90 
 
 86 
 
 58 
 
 78 
 
 85 
 
 53 
 
 80 84 
 
 19 
 
 88 
 
 45 
 
 95 
 
 67 
 
 80 55 83 61 76 60 77 91 87 78 77 60 65 90 
 
 86 
 
 59 
 
 81 
 
 84 
 
 7i 
 
 79 85 
 
 20 
 
 95 
 
 40 
 
 90 
 
 58 
 
 55 50 62 60 50 58 65 80 72 50 45 55 55 58 
 
 70 
 
 48 
 
 55 
 
 58 
 
 75 
 
 85 55 
 
 21 
 
 85 
 
 45 
 
 95 
 
 65 
 
 68 60 92 55 48 51 75 90 73 64 69 53 66 80 
 
 81 
 
 7i 
 
 87 
 
 78 
 
 62 
 
 70 83 
 
 22 
 
 81 
 
 48 
 
 95 
 
 67 
 
 73 55 92 72 66 50 75 87 70 62 70 66 64 74 
 
 78 
 
 53 
 
 80 
 
 75 
 
 68 
 
 82 85 
 
 23 
 
 76 
 
 50 
 
 95 
 
 65 
 
 64 57 92 61 59 74 58 82 79 66 68 63 60 84 
 
 83 
 
 62 
 
 80 
 
 71 
 
 69 
 
 75 87 
 
 24 
 
 77 
 
 45 
 
 95 
 
 64 
 
 62 60 93 66 59 67 69 83 80 63 68 56 75 85 
 
 89 
 
 61 
 
 87 
 
 72 
 
 70 
 
 65 91 
 
 1 Mr. Maurice J. Lacey did the statistical work on the Narration Scale. 
 
 2 Reader No. 4 did not read this set of themes. 
 
HARVARD-NEWTON BULLETIN 81 
 
 Table IV shows the ranking of each composition in narration 
 by each reader. 
 
 TABLE IV 
 
 Choice in Rank 
 fc P4 i 2 3 4 5 67 8 Q 10 ii 12 13 14 15 16 17 18 ig 20 21 22 23 24 25 
 
 I 
 
 2 
 
 3 
 4! 
 
 3 
 3 
 3 
 
 7 
 
 i 
 i 
 
 19 
 
 12 
 
 7 
 
 25 
 21 
 12 
 
 12 
 
 25 
 22 
 
 18 
 
 19 
 
 24 
 
 IS 
 
 22 
 25 
 
 21 
 II 
 
 18 
 
 22 
 
 18 
 
 17 
 
 z 
 IS 
 II 
 
 13 
 17 
 
 21 
 
 24 
 
 4 
 19 
 
 IO 
 24 
 
 16 
 
 4 
 13 
 
 23 
 
 5 
 14 
 IS 
 
 U 
 23 
 6 
 
 23 
 20 
 10 
 
 ii 
 
 16 
 9 
 
 20 
 
 5 
 5 
 
 6 
 
 IO 
 
 4 
 
 8 
 9 
 14 
 
 9 
 
 2 
 13 
 
 16 
 8 
 
 20 
 
 17 2 
 
 7 6 
 2 8 
 
 S 
 
 3 
 
 25 
 
 19 
 
 21 
 
 13 
 
 16 
 
 22 
 
 12 
 
 23 
 
 24 
 
 4 
 
 I 
 
 IS 
 
 5 
 
 II 
 
 7 
 
 10 
 
 17 
 
 14 
 
 18 
 
 2O 
 
 6 
 
 2 
 
 8 9 
 
 6 
 
 12 
 
 3 
 
 25 
 
 22 
 
 16 
 
 5 
 
 14 
 
 15 
 
 I 
 
 9 
 
 13 
 
 24 
 
 23 
 
 19 
 
 7 
 
 17 
 
 18 
 
 20 
 
 4 
 
 6 
 
 21 
 
 8 
 
 II 
 
 10 2 
 
 7 
 
 3 
 
 7 
 
 19 
 
 25 
 
 21 
 
 12 
 
 18 
 
 I 
 
 5 
 
 22 
 
 IS 
 
 13 
 
 14 
 
 24 
 
 8 
 
 9 
 
 23 
 
 4 
 
 17 
 
 ii 
 
 10 
 
 16 
 
 20 
 
 6 2 
 
 8 
 
 3 
 
 7 
 
 25 
 
 13 
 
 I 
 
 9 
 
 21 
 
 S 
 
 4 
 
 19 
 
 18 
 
 22 
 
 12 
 
 14 
 
 10 
 
 20 
 
 23 
 
 17 
 
 24 
 
 ii 
 
 6 
 
 16 
 
 2 
 
 IS .. 
 
 9 
 
 3 
 
 25 
 
 12 
 
 19 
 
 7 
 
 13 
 
 18 
 
 17 
 
 4 
 
 21 
 
 8 
 
 II 
 
 9 
 
 10 
 
 I 
 
 22 
 
 23 
 
 16 
 
 24 
 
 U 
 
 5 
 
 6 
 
 2O 
 
 IS 2 
 
 10 
 
 3 
 
 25 
 
 12 
 
 7 
 
 19 
 
 18 
 
 13 
 
 I 
 
 23 
 
 4 
 
 21 
 
 17 
 
 ii 
 
 8 
 
 9 
 
 22 
 
 16 
 
 S 
 
 10 
 
 24 
 
 M 
 
 6 
 
 20 
 
 IS 2 
 
 ii 
 
 7 
 
 3 
 
 5 
 
 21 
 
 25 
 
 12 
 
 19 
 
 18 
 
 I 
 
 24 
 
 22 
 
 9 
 
 14 
 
 ii 
 
 13 
 
 17 
 
 8 
 
 23 
 
 4 
 
 6 
 
 10 
 
 20 
 
 IS 
 
 16 2 
 
 12 
 
 3 
 
 12 
 
 13 
 
 I 
 
 25 
 
 18 
 
 4 
 
 ii 
 
 23 
 
 IS 
 
 21 
 
 19 
 
 I? 
 
 22 
 
 7 
 
 5 
 
 16 
 
 14 
 
 24 
 
 8 
 
 10 
 
 9 
 
 20 
 
 6 2 
 
 13 
 
 25 
 
 3 
 
 7 
 
 19 
 
 13 
 
 15 
 
 ii 
 
 10 
 
 12 
 
 21 
 
 24 
 
 23 
 
 22 
 
 I 
 
 4 
 
 8 
 
 S 
 
 6 
 
 9 
 
 14 
 
 18 
 
 17 
 
 20 
 
 2 16 
 
 14 
 
 3 
 
 19 
 
 25 
 
 7 
 
 12 
 
 15 
 
 i 
 
 S 
 
 13 
 
 21 
 
 23 
 
 18 
 
 22 
 
 24 
 
 17 
 
 20 
 
 16 
 
 ii 
 
 14 
 
 4 
 
 6 
 
 10 
 
 8 
 
 9 2 
 
 IS 
 
 3 
 
 7 
 
 25 
 
 21 
 
 19 
 
 12 
 
 18 
 
 15 
 
 9 
 
 I 
 
 13 
 
 17 
 
 22 
 
 8 
 
 5 
 
 II 
 
 14 
 
 23 
 
 10 
 
 24 
 
 6 
 
 4 
 
 16 
 
 20 2 
 
 16 
 
 25 
 
 3 
 
 7 
 
 12 
 
 21 
 
 19 
 
 18 
 
 I 
 
 22 
 
 13 
 
 24 
 
 15 
 
 23 
 
 14 
 
 S 
 
 17 
 
 8 
 
 10 
 
 ii 
 
 9 
 
 16 
 
 4 
 
 20 
 
 6 2 
 
 i? 
 
 3 
 
 18 
 
 19 
 
 I 
 
 22 
 
 21 
 
 24 
 
 25 
 
 S 
 
 12 
 
 II 
 
 IS 
 
 14 
 
 13 
 
 7 
 
 10 
 
 20 
 
 9 
 
 23 
 
 17 
 
 8 
 
 6 
 
 16 
 
 4 2 
 
 18 
 
 3 
 
 12 
 
 18 
 
 7 
 
 I 
 
 19 
 
 22 
 
 25 
 
 13 
 
 24 
 
 21 
 
 ii 
 
 15 
 
 U 
 
 5 
 
 9 
 
 4 
 
 17 
 
 10 
 
 8 
 
 20 
 
 6 
 
 23 
 
 16 2 
 
 19 
 
 3 
 
 12 
 
 18 
 
 i 
 
 13 
 
 19 
 
 25 
 
 22 
 
 7 
 
 21 
 
 5 
 
 24 
 
 U 
 
 IS 
 
 9 
 
 ii 
 
 23 
 
 4 
 
 17 
 
 8 
 
 16 
 
 10 
 
 20 
 
 6 2 
 
 20 
 
 I 
 
 3 
 
 24 
 
 12 
 
 23 
 
 13 
 
 19 
 
 II 
 
 7 
 
 8 
 
 4 
 
 10 
 
 18 
 
 22 
 
 25 
 
 21 
 
 17 
 
 16 
 
 S 
 
 14 
 
 9 
 
 6 
 
 20 
 
 IS 2 
 
 21 
 
 3 
 
 7 
 
 12 
 
 21 
 
 I 
 
 25 
 
 19 
 
 18 
 
 22 
 
 II 
 
 13 
 
 20 
 
 24 
 
 15 
 
 5 
 
 17 
 
 4 
 
 M 
 
 23 
 
 6 
 
 8 
 
 16 
 
 10 
 
 9 2 
 
 22 
 
 3 
 
 7 
 
 12 
 
 25 
 
 24 
 
 I 
 
 21 
 
 19 
 
 II 
 
 22 
 
 18 
 
 5 
 
 8 
 
 13 
 
 15 
 
 23 
 
 4 
 
 9 
 
 16 
 
 17 
 
 14 
 
 6 
 
 20 
 
 IO 2 
 
 23 
 
 3 
 
 7 
 
 25 
 
 18 
 
 19 
 
 12 
 
 21 
 
 13 
 
 I 
 
 24 
 
 10 
 
 22 
 
 23 
 
 IS 
 
 14 
 
 4 
 
 5 
 
 16 
 
 20 
 
 8 
 
 17 
 
 9 
 
 II 
 
 6 2 
 
 24 
 
 3 
 
 7 
 
 25 
 
 19 
 
 21 
 
 18 
 
 12 
 
 13 
 
 I 
 
 17 
 
 22 
 
 23 
 
 i 
 
 15 
 
 10 
 
 8 
 
 24 
 
 4 
 
 14 
 
 5 
 
 20 
 
 6 
 
 9 
 
 16 2 
 
 Reader No. 4 did not read this set of compositions. 
 
82 HARVARD-NEWTON BULLETIN 
 
 THE COMPLETED NARRATION SCALE 
 
 No. i. "A" GRADE COMPOSITION. VALUE, 93.5% 
 
 MY FIRST RIDING LESSON x 
 
 I gazed down into the riding ring from the balcony where I 
 was sitting. The riding master was cantering around the 
 ring with a young girl who seemed to be an experienced rider. 
 Riding looked easy. 
 
 5 After another round he rode up to the doors which a man 
 outside flung open, and rode out. A few minutes later he 
 called up the stairs, " I am ready for you, Miss Speare," and 
 taking my crop I rushed eagerly down the stairs, for this was 
 my first riding lesson. 
 
 10 I scrambled on to the back of a small horse that stood pa- 
 tiently quiet, waiting. 
 
 " Are your stirrups all right ? " the riding master asked, 
 when I was on. 
 
 " Yes," I replied, "They are just the right length." 
 15 " Thatls good," and lie leaped into the saddle and we rode 
 into the ring. 
 
 " You have not ridden before, have you ? " he asked as we 
 walked slowly around the ring. 
 
 " No," I replied, " I have n't." " But I was thinking to 
 20 myself how easy riding is ! 
 
 " We will trot now," he said, after a little and the horses 
 began to trot slowly. 
 
 Then I discovered my mistake. Riding is not especially 
 easy for beginers. And trotting is not as easy as walking. 
 25 " Rise with the horse," he commanded. " Cling on with 
 your knees and rise in your stirrups." 
 
 Easier said then done. I bounced up and down until I 
 thought I should fall off. It looked so easy when other 
 people trotted! He seemed to do it so easily, why couldn't 
 30 I ? Yet I could n't and that was plain. 
 
 1 Composition No. 3 in the complete set. 
 
HARVARD-NEWTON BULLETIN 83 
 
 " I c-can't s-seem to r-rise with h-him. H-he b-bounces 
 m-me awful," I gasped finally. He pulled in and we walked 
 for a while. 
 
 " Now try to rise in your stirrups," he said as we began to 
 trot again. 35 
 
 I got along better this time although I still bumped 
 terribly. 
 
 Then he discovered that the hour was up and we rode out. 
 He dismounted and helped me off and when I reached the 
 ground I felt very unsteady. 40 
 
 I staggered up the stairs, for I was very wobbly, seized my 
 crop and hurried down. If riding was like this all the time! 
 He was talking to mother. 
 
 " All right," I heard him say, " A week from to-day; same 
 time. She did pretty well for the first time. Too bad she is 45 
 so stiff." 
 
 I didn't feel like another lesson at any time but I didn't 
 say so and with mother I left the riding school and my first 
 riding lesson was at an end. 
 
 Merits 
 
 The situation is vividly portrayed, with plenty of conversation. From 
 the lively introduction to the direct conclusion, the interest is kept up by 
 well-selected details. The paragraphing is good, and the mastery of quo- 
 tation marks is evident. Good choice of words is shown in such natural 
 phrasing as rushed, gazed, flung, scramble, bounced, staggered. The connec- 
 tion between the paragraphs is unusually good. 
 
 Defects 
 
 The defects are in matters of form. There is unnecessary repetition of 
 the word ride. Punctuation is omitted in several places. There are mis- 
 takes in spelling (lines 10, 24, 27). In that's good, the apostrophe is omitted. 
 The concluding sentence is loosely constructed. 
 
 Comparison 
 
 This theme holds a higher place than No. 2 because of its excellent choice 
 of details and its vivid expression. It is a more ambitious piece of work, 
 and the interest is more surely sustained. 
 
84 HARVARD-NEWTON BULLETIN 
 
 No. 2. " B " GRADE COMPOSITION. VALUE, 83.0% 
 MY TRIP TO FAIRYLAND l 
 
 It was about four o'clock in the afternoon, and as I was 
 very tired I sat down by the fire to rest. Suddenly I heard a 
 small voice calling me by name. I immediately looked up 
 and beside me I saw what I thought was a fairy. I was very 
 5 much excited as I had never seen a fairy but had only read of 
 them. 
 
 " Would you like to come with me to Fairyland and rid 
 yourself of your cares for a while ? " she questioned. 
 
 " I would love to," I answered. 
 10 " Then only follow me," she said. 
 
 So out of the room we went and down the cellar stairs. 
 
 " Where are you leading me ? " I asked. 
 
 " You shall soon find out," she answered. 
 
 We went into the darkest corner of the cellar. She showed 
 15 me a small trap door which she went through, leaving me 
 alone. 
 
 " But how am I to get through ? " I asked. 
 
 Without a reply she touched me with her small wand, and 
 before I had a chance to say a word I found myself through 
 20 the door and in a beautiful garden which was filled with fra- 
 grant flowers of many kinds. She led me through it, and at 
 one end of it I saw a white marble palace. 
 
 " This is my home," she said. Do you care to enter ? " 
 
 " That would be delightful," I said. 
 
 25 Up the stairs we went and down a long hall. Finally we 
 came upon a large beautiful room. It was filled with danc- 
 ing fairies who were having a gay time. My fairy introduced 
 me to some of the fairies, and then joined in the merry mak- 
 ing. I stood and looked at the splendor of it all, when sud- 
 30 denly I heard a voice calling me, and I looked up. 
 
 It was my mother who said, " I have called you three 
 times. You must have been asleep." 
 
 1 Composition No. 19 in the complete set. 
 
HARVARD-NEWTON BULLETIN 85 
 
 Merits 
 
 The spelling, punctuation, sentence structure, and especially the para- 
 graphing of this composition, are accurate. The story, as a whole, has 
 action; the conversation is well handled; there is a studious effort to avoid 
 repetitions, as, questioned, asked, answered, said (lines 8-12). 
 
 Defects 
 
 The chief defect is a childish use of tiresome short sentences. The verb 
 forms do not produce action enough. In line 23 quotation marks are 
 omitted before Do, and there should be a hyphen in merry-making (line 28). 
 
 Comparison 
 
 The superior paragraph structure of this theme is sufficient to place it 
 above No. 3. It lacks the merit seen in the easy and spirited conversation 
 of No. i. 
 
86 HARVARD-NEWTON BULLETIN 
 
 No. 3. "C" GRADE COMPOSITION. VALUE, 76.2% 
 LOST IN THE WOODS OF CONNECTICUT 1 
 
 " Who 's going blackberrying to-day ? " cried my cousin 
 Margaret waving her pail. 
 
 " I am," I said coming out of the house a pail in one hand 
 and a wide brimmed hat in othe other. 
 5 " Allright," answered Margaret, " but where's Julia ? " 
 
 " Must be in the house," I replied. 
 
 We soon found her diligently reading a book and after some 
 urging she decided to go. 
 
 It was about eight o'clock when we started for if we 
 10 waited till later on in the day it would be to hot. 
 
 We walked a long ways to the other end of the farm after 
 climbing stonewalls and having several encounters with 
 snakes. 
 
 " How much farther do we have to go ? " I cried growing 
 15 impatient. 
 
 " The other side of this field," Margaret called back. 
 
 After climbing the stonewall we came in sight of big 
 stretches of blackberries. 
 
 We fell to work right away and picked very fast. 
 20 In a little while we stopped and found we had worked a 
 good ways. As we went on the berries seemed to grow big- 
 ger and sweeter. 
 
 After filling my pail I found Margaret was through long 
 before me and was sitting in their midst eating. We waited 
 25 till Julia was through and then started for home. We were 
 so busy talking Margaret did n't notice which path we took 
 and after walking a longs ways she discovered our mistake. 
 I wanted to turn back and go the way we came but she said 
 it would take to long so we struck out into the woods. The 
 30 farther we went the more perplexed we were till finally Mar- 
 garet sat down under a tree in despair. 
 
 1 Composition No. 22 in the complete set. 
 
HARVARD-NEWTON BULLETIN 87 
 
 This would never do so I decided to climb a nearby tree to 
 see if we could see any signs of a house. 
 
 With much pulling and tugging I got up the tree and saw a 
 house to the right. We reached the house tired and hungry. 35 
 
 We were lucky to come to come to such a place for the mis- 
 tress proved to be very good. After resting one of the hired 
 men drove us home just in time to see a searching party going 
 out to look for us as it was now late in the afternoon. 
 
 Merits 
 
 The most commendable feature of this narrative is the management of 
 the conversation. This is accurately punctuated and contains a variety 
 in the verbs of saying. Placed at the beginning, it gives an element of 
 spontaneity and arouses the immediate interest of the reader. There are 
 several well-chosen expressions such as several encounters with snakes, 
 pulling and tugging, and in despair. The arrangement of ideas is logical. 
 
 Defects 
 
 Except for the incompatibility between the content and the title, the 
 defects are mechanical; (i) poor paragraphing except in the conversation; 
 (2) the unskillful use of the compound sentence in the sentences beginning 
 on lines 7, 9, 25; (3) the unemphatic position of the participial phrase in 
 lines 12 and 13, and of the dependent clause in the last sentence; (4) the 
 absence of commas after introductory clauses and phrases, in some instances 
 causing an ambiguity of meaning as hi sentences commencing in lines 21 
 and 37; (5) the grammatical error in the incoherent reference of the pro- 
 noun their in line 24; (6) the evident carelessness in the spelling of the, all 
 right and too and in the repetition of to come in line 36; and (7) the poor 
 choice of words in the colloquial nearby and in the repeated use of long ways. 
 
 Comparison 
 
 This theme is superior to No. 4 in its maturity of thought and phrasing; 
 it is inferior to No. 2 in mechanical form. 
 
88 HARVARD-NEWTON BULLETIN 
 
 No. 4. " D " GRADE COMPOSITION. VALUE, 66.4% 
 A HAPPY BUT EXCITING AFTERNOON l 
 
 Allen had just gotten a new motor-boat for his birthday 
 and wanted to try it out as it was Saturday. " Well," he 
 said, " will you come with me Emily. You can bring Edna 
 if you wish." 
 
 5 "All right," said Emily, and off she went to get Edna, her 
 cousin. 
 
 They started out at Marblehead and was going to go as far 
 out in the harbor as they could. 
 
 All went well for about one hour, then it was getting rough. 
 10 The girls got frightened, but, they had more to be frightened 
 about later, the propeller broke off. All there was to do was 
 to go where ever the waves would take them. 
 
 They were heading for open sea. " Oh," said Edna, " how 
 perfectly horrible." 
 
 15 " I 'm not coming out in this motor-boat again," said 
 Emily. 
 
 " I guess you won't, Emily, but look, here comes a ferry- 
 boat." 
 
 So it was, " now for the fun," said Allen, who was not the 
 20 least bit excited, " I 'm going to call their attention." 
 
 The ship came nearer and nearer to them. Soon it was 
 within reach and Allen made a grab for it. He got hold of 
 the side and helped the girls in. Then he got a rope and tide 
 it to the motor-boat. 
 
 25 " Well," said the girls to-gether, " we will never go in a 
 motor-boat with you again. 
 
 Merits 
 
 The merits are two: (i) the child had a fairly well planned story in mind; 
 (2) he has an accurate knowledge of the use of quotation marks. 
 
 1 Composition No. 14 in the complete set. 
 
HARVARD-NEWTON BULLETIN 89 
 
 Defects 
 
 The defects are as follows: (i) the paragraphing is poor; (2) there are 
 several examples of the " run-on " sentence (lines 9, n, and 20); (3) the 
 words propeller, wherever, together, and tied are misspelled; (4) can is used 
 in line 3 instead of may, (5) in line 7 there is a singular verb used with a 
 plural subject; (6) there are errors in punctuation (lines 3, 10, 14, and 19); 
 (7) the vocabulary is childish. 
 
 Comparison 
 
 This theme follows No. 3 in the scale because of the inferior development 
 of its plot and the greater immaturity of its expression. 
 
90 HARVARD-NEWTON BULLETIN 
 
 No. 5. " E " GRADE COMPOSITION. VALUE, 56.0% 
 A GHOST STORY 1 
 
 " Come Joe, run faster we must get Harry to come with us 
 and we have but few minutes to do it in if we want to get 
 there before dark/' said John as he leaped a fence and started 
 toward a house in the distance. 
 
 5 Joe jumped the fence and ran after John as fast as he could 
 in order to catch up with him so they could go in together. 
 
 They reached the house all out of breath. 
 
 " Hullo boys! where are you going, can I be one com- 
 pany ? " were the series of questions poured forth to John and 
 10 Joe as they entered the house. 
 
 John and Joe told them where they were going, and Harry 
 readly consented to go. 
 
 They reached the house and ate their supper and went to 
 bed. 
 
 1 5 They were not asleep long when Joe woke up and hollered 
 for John and Harry to come to him because he saw a ghost. 
 
 Harry jumped up and entered the room in which Joe was 
 sleeping. 
 
 " Where is the ghost," he asked, " I don't see any." 
 20 He is gone now, said Joe. In the morning they thought 
 they would investigate. 
 
 They found it was only a white post and a white sheet be- 
 hind it. They then started home and told their parents about 
 their adventures. 
 
 Merits 
 
 The writer has a strong and dramatic opening, and throughout the theme 
 handles his dialogue well. His sense of order is good and his vocabulary 
 adequate. Such expressions as the series of questions poured forth and readily 
 consented are good. 
 
 Defects 
 
 The good promise of the beginning is not fulfilled; the climax is not 
 worked up successfully and the whole trails off into a pointless incident. The 
 
 1 Composition No. 6 in the complete set. 
 
HARVARD-NEWTON BULLETIN 91 
 
 mechanical errors are numerous; almost every sentence is made a separate 
 paragraph; there are two examples of the " run-on " sentence (lines i and 
 19); there are several comma omissions and one use of a comma where 
 there should be a question mark (line 19) ; there is an error in tense sequence 
 (line 1 5) ; the words of the are omitted in line 8 ; readily is misspelled (line 1 2) . 
 
 Comparison 
 
 While more dramatic than No. 4, the theme is properly placed lower on 
 account of the large number of mechanical errors and the drooping narrative 
 interest. 
 
92 HARVARD-NEWTON BULLETIN 
 
 No. 6. " F " GRADE COMPOSITION. VALUE, 46.9% 
 LIFE IN A CABIN l 
 
 In a woods our cabin is which is made of logs. In side are 
 two rooms furnished with rough hand made furniture. On 
 the wall of the den hung skins and many pictures. 
 
 We had to have something to eat so we had to find a river 
 5 where we could fish. We got enough fish and soon they were 
 over the fire frying. 
 
 Next day an Indian came with some meat. We gave him a 
 few coppers had he was real pleased. 
 
 " Jingle jingle," was the noise out side as the horse gal- 
 10 loped. It was cousen Joe. 
 
 " Hurrah ! now for some fun," I cried. " Come toboggan- 
 ing ? See that that hill over there well its great toboggan- 
 ing." 
 
 " Come in," said Joe, " till I get my moccasins ? " 
 15 What fun we had many tumbles we got. Just as we were 
 going home we saw a deer laying in the snow. We crept up 
 to him but he did not move. Than we put him on the tobog- 
 gan and took him home. He was cold but we gave him 
 some warm milk and he tryed to get up, But he could not. 
 20 " Poor thing he must be hurt," said Joe. " His leg is 
 broke don't you see." Then we bathed his leg and bound it. 
 Soon he was able to stand on his legs and became a house 
 pet. 
 
 " Oh, dear its snowing," said Joe, " Lets not go out as my 
 25 snoweshoes are not here." 
 
 Merits 
 
 The merits of this theme are: (i) an attempt to make an introductory 
 paragraph; (2) the introduction of conversation; (3) the use of interrogative 
 and exclamatory sentences; (4) the simple and natural language; (5) the 
 orderly sequence of points in the story. 
 
 1 Composition No. 2 in the complete set. 
 
HARVARD-NEWTON BULLETIN 93 
 
 Defects 
 
 Though well begun, after the first paragraph this theme is simply a series 
 of disconnected events and apparently unrelated statements, (i) Punctua- 
 tion is faulty; (2) sentences are badly constructed ; (3) the meaning in 
 several places is not clear; (4) there are many mechanical errors, such as 
 than for then', broke for broken; laying for lying; had for and; (5) the 
 apostrophe is omitted in contractions ; and (6) the tense changes abruptly 
 in the opening paragraph. 
 
 Comparison 
 
 This theme is inferior to No. 5 in sentence structure, in connected arrange- 
 ment of material, and in mechanical accuracy. 
 
UNIVERSITY OF CALIFORNIA LIBRARY