ANALYSIS AND MEASUREMENT OF SPELLERS BY ARLIE GLENDi CAPPS B. S. University of Missouri, 1916 A. M. University of Missouri, 1917 THESIS Submitted in Partial Fulfillment of the Requirements for the Degree of DOCTOR OF PHILOSOPHY IN EDUCATION IN THE GRADUATE SCHOOL OF THE UNIVERSITY OF ILLINOIS 1921 ■ jo >V . . . i TABLE OF CONTENTS Chapter Page I. Introduction Statement of Project 1 Need for the Project 3 Limitations of Project 6 Earlier Investigations 7 Score Cards 10 Bases of the Project 13 II. Procedure in Analyzing the Spellers Spellers Analyzed 19 Procedure in Analyzing the Spellers 19 III. Extent of Vocabulary Extent of Vocabulary in Each Book 26 Extent of the Vocabulary in Each Grade 31 IV. Selection of Vocabulary Selection of Vocabulary by Books 37 Selection of Vocabulary by Grades 44 V. Gradation of Vocabulary 56 Median Difficulty of Each Grade in Each Speller 60 Difficulty by Grades 72 Median Difficulty of Each Speller 8o Comparison of the Books on the Basis of Difficulty 84 VI. Repetition of Words 91 Average Number of Repetitions of All Words by Grades 94 Average Repetitions of Ayres-Teachers College Words 99 Repetition of Words from Grade to Grade 100 ■ . . . 1 7T|1 3-1 Chapter Page Repetition of a Selected List of Words 103 Correlation Between Repetition and Difficulty of Words 106 Correlation Between Repetition and Difficulty by Grades in Each Book 110 Correlation Between Repetition in Spellers and in Adult Use 113 VII. Forms of Presentation of Words Percent of Spellers Consisting of Dictation Material 120 Percent of All Material in Each Grade Consisting of Dictation Material 121 Grade Norms for Certain Forms of Presentation Based on Present Practice 124 Collation of All Forms of Presentation Examined 126 VIII. Two Dimensional Check Lists 132 Check Lists 142 IX. Summary and Suggestions Summary 1 49 Suggestions 154 Selected Bibliography Digitized by the Internet Archive in 2015 https://archive.org/details/analysismeasuremOOcarp iii LIST OF TABLES Table Page I. Spelling textbooks studied 20 II. Number of different words in each book 27 III. Number of different words in each book (Woody) 30 IV. Number of different words in each grade 32 V. Extent to which the words of the Ayres-Teachere College list are found in the textbooks 39 VI. Rank of spellers on percent of Ayres-Teachers College words contained in their vocabularies 43 VII. Number of Ayres-Teachers College words per grade 46 VIII. Average and hypothetical number of Ayres-Teachers College words per grade 4 7 IX. Percent of words in each grade that are Ayres- Teachers College words 51 X. Percent of Ayres-Teachers College words in each grade 53 XI. Median percent of each of the components of the Ayres-Teachers College list in each grade 55 XII. Median difficulty of the Ayres-Teachers College words as found in each grade and book 61 XIII. Median difficulty of the Ayres words as found in each grade and book 65 XIV. Median difficulty of the Teachers College words as found in each grade and book 68 XIV-A. The median difficulty of the Teachers College words as found in each grade and speller expressed in terms of columns and words on the Ayres Scale 71 ' 1 • • 4 . . ' ! iv Table Page XV. Index numbers by grades, Ayres-Teachers College words 73 XVI. Index numbers by grades, Teachers College words 75 XVII. Grade intervals after Buckingham 77 XVIII. Median difficulty of each book. First method 8l XIX. Median difficulty of each book. Second method 82 XX. Median difficulty of each book. Third method 83 XX-A. Comparison of median difficulty with median pupil ability by grades 87 XXI. Rank of spellers based on adjustment to the ability of pupils 90 XXII. Number of words, number of occurrences, and average occurrences per word 93 XXIII. Average number of repetitions per word 95 XXIII-A. Range in average repetitions per word for all words 96 XXIV. Average number of repetitions per word for the Ayres-Teachers College words in each grade 98 XXV. Repetition of words from grade to grade 101 XXVI. Average number of occurrences per word for selected list of 100 104 XXVII. Behavior of the easy word "some" 104a XXVIII. Behavior of the median difficulty word "write” 104b XXIX. Behavior of the difficult word "sincerely" 104c XXX. Correlations between repetition and difficulty based on words common to each speller and to the Ayres-Teachers College list 107 XXXI. Correlations between repetition and difficulty based on words common to each speller and to the Teachers College list 108 . , .. • • - - . . - * * ' . * * • . . . - * * ■ . , V Table Page XXXII. Correlation between the repetition of the words in 109 the speller and their difficulty XXXIII. Correlation between the repetition of the words in the speller and their difficulty 111 XXXIV. Correlations between repetition in speller and in life based on 100 words selected at random from the Ayres- Teachers College list and common to each of the spellers 115 XXXV. Percent of spellers consisting of dictation material 120 XXXVI. Amount of dictation expressed as percent that the words printed in dictation are of all the printed words 122 XXXVII. Grade norms for dictation material based on medians of eight spellers 123 XXXVIII. Percent certain forms of presentation is of all forms of presentation 126 XXXIX. Number of printed words in each form of presentation 128 XL. Percent that the number of printed words in each form of presentation is of all the printed words in the book 130 XLI. Data obtained by the series of two dimensional check lists 135 XLII. Conspectus of rankings of spellers 139 XLIII . Norms derived from present practice 141 . - . - ■ . . . Vi LIST OF FIGURES Figure Page 1 Number of different words by grades in typical spellers 34 2 Range of median grade difficulties for ten spellers based on Ayres-Teachers College words 63 3 Range of median grade difficulties for ten spellers based on Ayres words 67 4 Range of median grade difficulties for ten spellers based on Teachers College words 70 5 Grade intervals between index numbers 76 6 Ability and difficulty intervals 79 7 Composite grade difficulty in relation to the theoretical difficulty and to the grade ability 88 8 Percent of spellers composed of dictation material in typical spellers 125 vii ACKNOWLEDGEMENTS This dissertation was made possible by the efforts of many people to whom I am greatly indebted. My greatest indebtedness is to Dr. B. R. Buckingham who gave assistance by encouragement, by technical advice, and by providing cler- ical help to do a major portion of the large amount of mechanical labor involved in carrying on the various activities of the project. To Dean C. E. Chadsey, I am grateful for the opportunity which he gave me to continue the project over an extended period of time and for encouragement and counsel. The criticisms of Drs. E. H. Cameron and Walter S. Monroe were especially valuable and thoroughly appreciated. Finally, I wish to express my appreciation of the faithfulness and the carefulness of my wife, of the clerks who spent countless hours on uninteresting routine, and of the labors of many investigators whose findings have helped to make possible this project. ANALYSIS AND MEASUREMENT 0 F SPELLER S CHAPTER I INTRODUCTION Statement of the Project Within the last decade long strides have been made in the field of measuring educational products. Tests and scales have been devised for a major- ity of the subjects in the elementary and secondary school curriculums. The extent of the use of these instruments for measuring the educational product may be illustrated by the fact that one of the many agencies for their distribution in this country — the Bureau of Educational Research, College of Education, Uni- versity of Illinois— distributed last year over two and one half million copies. While, however, educational scientists have gone a long way towards developing the measurement of the results of instruction, they have done comparatively lit- tle in measuring the instruments of instruction. The most important instrument of instruction is the textbook. As far as possible, the same accurate informa- tion concerning them should *be had as we now have concerning the product of our schools. In the science of engineering and in any of the laboratory sciences the expert worker has accurate and detailed knowledge of his instruments . He knows to a high degree of precision the accuracy of his balances, thermometers, and calipers. The reliability of these instruments is expressed in terms of well-known error constants. In short, the instruments are accurately calibrated. Similarly, the instruments of instruction should be calibrated. A complete analogy with the measuring devices of the physical sciences - 2 - would demand that we know concerning the textbooks not only their contents and the methods of teaching which they exemplify but also the precise effect in school procedure which they produce. Here, as elsewhere in educational measure- ment, the analogy with the physical sciences is imperfect. The degree of accura- cy of a spring balance can never be attained in the field of human behavior. This, however, does not justify us in neglecting to secure such accuracy as we can. In the case of textbooks we may neither know now, nor ever be able to find out the precise effect which a given content and method may produce in a given individual. However, we may find out first what the contents of the book are, second what relation these contents bear to the needs of life, and third what opportunity the book affords for the mastery of these contents. What use is made of the offerings of the textbook — what ability, experience, training, and effort, what ideals and attitudes may be possessed by the teacher and the pupil— must be sought outside the covers of the textbook. These types of inquiry await the attention of investigators. It is certain, however, that such larger problems may be attacked with far higher prospects of success if the groundwork is laid in an accurate evaluation of the textbook itself. Accordingly, we should know the contents and the methods used in the instruments of instruction, the text- books, as accurately and in as great detail as the science of education can determine. Therefore, the original purpose of this project was to examine a num- ber of different types of textbooks, but it was found after careful examination of the literature that detailed subordinate data did not exist for any of the textbooks except spellers. By "subordinate data" we mean in part reliable infor- mation as to the demands made upon the schools by society. For example, what subject matter— what knowledge, habits, and appreciations— should the pupil gain by his school experience in a given subject in order that he may be able to meet the needs that his social life forces upon him? In the subordinate data we are, also, looking for reliable information as to the ability of the pupil to handle - 3 - the subject matter that he will have presented to hirn in order that he may be able to meet the demands of society. For example, what knowledge is he able to grasp and use, what habits are within his ability to develop, and what appreci- ations is he able to acquire? In our project we therefore sought trustworthy data as to the content demanded by society and as to the gradation of this con- tent in order that the pupil might grasp it most efficiently. These data were necessary before the contents and the methods of textbooks could be examined and evaluated. They did not exist to a sufficient degree for our purposes in any subject except spelling. Consequently, this dissertation is concerned with a project in which the purpose was to analyze and to measure in so far as possible the contents and the methods of spelling textbooks. While we sense the crudeness of the technique and the meagerness of the results for this project, it is hoped that it will con- stitute the first of a series of similar projects in the analysis and measurement of the instruments of instruction which will be made as rapidly as preliminary investigations will permit. Meed for the Present Project That there is a felt need on the part of the leaders in education for projects in analyzing and measuring textbooks, that is the instruments of in- struction, is well expressed by at least three of our contemporary schoolmen. The following quotation from a recent address given by Dr. E. L. Thorndike while dealing specifically with arithmetics might with equal propri- ety have dealt with spellers. "The general text, indeed, of what I have to say will be this: Nei- ther you, nor I, nor anyone I know of really knows the instruments that we work with. We are like physicians who pour drugs out of a bottle not knowing what is in it. "Evidently we require, for knowledge of the instruments of instruction, the same careful measuring and counting and full analysis which we have learned to apply to the examination of children’s achievements. "It is a question of some consequence to know how much habituation is required to inculcate certain fundamental facts in arithmetic. We cannot, with- out having somebody follow the child through school and keep track of all the oral drills, making discounts for times he is inattentive and the like, get meas- ures of children's actual experiences. We can, however, as I have done, count up the number of experiences the child would have if he did honestly all the work outlined in some standard educational instrument. "In general, the neglect of childish interests does not seem to be due to provision for some other end, but to the same inertia of tradition which has carried over the problems of laying walls and digging wells into city schools, whose children never saw a stone wall laid or a well dug. "I hope I have convinced you that we can apply scientific methods to measuring the instruments of instruction with the same hope of benefit that we have received from its application to measurements of school achievement ." 1 Dr. C. H. Judd well expressed the need for such a project as ours in the following quotation: )"There is no influence in American schools which does more to determine what is taught to pupils than does the text-book. Yet this important factor in our educational system has until recent years altogether escaped critical study. — -It is put into the hands of teachers who are for the most part so limited in experience and training that they never think of questioning the method suggested by the book and never dream of doubting the validity of the educational princi- ples on which the author's selection and arrangement of material are based. "Every text-book, whatever its source, has characteristics which can be accurately tabulated and described. Each book has a kind of personality which can be measured no lees accurately than the physical characteristics of a man. ^"Experience leads us to demand, therefore, some kind of analysis of books which will clarify the teacher's judgment. The text -bo ok must be analyzed by some impersonal method in order to prepare the way for the teacher's judgment. "Evidently analysis of text-books will help in the reforms which are coining in the interests of economy. Following Dr. Judd's article in succeeding issues of the Elementary School Journal were descriptions of a number of projects in analyzing textbooks, but none of them dealt with spellers. More recently Dr. Buckingham has called attention to the need for analyzing and measuring the instruments of instruction. He says: "The practice of teachers is largely controlled by the text-book. As is frequently pointed out, it plays a unique part in the affairs of the American * ^Thorndike, ~E. L.: "Recent developments in educational measurements, "Fifth Con- ference _ o^_Educational M e asurements , (Bulletin of the Extension Division, Indiana University, December, 1918.) Pp.^-?2. 2 Judd, C. H. "Analyzing text-bocks." Elementary School Journal, 19:143-154, October, 1918. " - 5 - school. It is, therefore, highly important that the textbook should exemplify the best that competent research has established. There should be an easy trans- fer of the findings of investigators to the content of school books. It is not enough to train the teacher in the spirit of research; the textbook should be the concrete fruition of research. •'Investigations looking toward the improvement of textbooks are under way in a number of places and it is certain that this type of investigation will become more prevalent when its importance is better understood. Titles of recent masters' and doctors' theses make clear the fact that in departments of education in colleges and universities, the two fields of research which are regarded as most significant are the study of teaching methods and the analysis of curricu- lums. In the public schools, however, it is equally clear that the teaching methods and the curriculum actually put into practice are the methods and the curriculum of the textbook. To be effective, therefore, it is in the textbooks that the new methods and curriculums must be expressed. "Although we are in the midst of researches whose results ought to be at once reflected in better textbooks, better textbooks are slow to appear. If it is a day of promise in research, it is also a day of stagnation in textbook making. The results, the methods, and the spirit of research will never become fully applicable in the work of the schools until in addition to their apprehen- sion by teachers, they find expression in books which are placed in the hands of pupils. Seven months later Dr. Buckingham again pointed out the need for pro- jects similar to the one concerned with in this dissertation. He says: "In the textbook business it is clear that research workers in educa- tion are needed. The critical attention now being given to the contents of school books discloses many ways in which they may be improved. Not long ago we heard a very able man, himself an eminent research worker, say after having examined minutely a large number of texts, that most books were adopted without anyone knowing what they contained— neither the committee that judged them nor the teachers who used them. Of course, he did not mean that the chief topics presented in the books were not known— although even that degree of knowledge is often lacking. What he meant v/as that the contents of the books were unanalyzed with reference to desirable objectives ." 4 The preceding quotations put the issue clearly. These writers are asking that projects in analyzing and measuring the textbooks by the modern tools of research and in the light of recent investigations be carried on. Our project is designed to meet their demands in one definite field, namely, that of spelling. 3 ' - 1 11 1 ■ Buckingham, B. R. "Textbooks: their cost and improvement" (Editorial). Jour- nal of Education al Research , 1:222-24, March, 1920. ^Buckingham, B. R. "Research in textbook publishing" (Editorial). Journal of Edu c ational Resea rch, 2:670-72, October, 1920. - 6 - Limitation s o£ the Pro j ect Our activities were limited to analyzing and measuring certain aspects of the subject matter and the methods of presentations found in ten recently copyrighted, popular spelling textbooks designed for the grades of the elementary schools. The aspects examined are: (a) the extent, the selection, and the grada- tion of the vocabulary; (b) the repetition of words; and (c) the following forms of presentation— column, column review, supplementary, dictation, dictation re- view, word building, word analysis, column homonyms, dictation homonyms, review homonyms, phonics, and plural formation. (For the detailed interpretation that we made of these forms, the reader should consult Chapter II.) From the foregoing statement of limitations it may be seen that in the first place we paid no attention to other forms of English than spelling. We made no study of the offerings in the spellers in the way of memory gems — how many or how much, source, adaptability to the interests, needs, and problems of the students, etc. We did not study punctuation marks, abbreviations, letter writing, etc. Many of these forms of English are found in the spellers and prop- erly so, especially if they are introduced for spelling purposes. However, we found it necessary to pass by this phase of the spellers. In the second place, we did not examine every aspect of spelling. We paid no attention to syllabication, diacritical marks, directions to teachers, directions to pupils on how to study the lessons, rules for spelling, prefixes and suffixes, provisions for enlarging and enriching the pupil's vocabulary, "blunder spots," forms of grouping, proper names, pictures, cost, quality of paper, etc. All the preceding items and many implied by "etc." should be taken into consideration in selecting a speller. Hence, it is seen that our study is not complete for this purpose. We are merely concerned with the word offerings — the extent, the selection, the gradation, and the repetition of words — and with - 7 - certain forms of presentation. Many of the items, however, with which our pro- ject is not concerned sire obvious. The real difficulty in judging a speller lies precisely in reaching a conclusion regarding the vocabulary — using vocabulary in the broad sense. However important other items may be, no item vies in importance with the one in which we are concerned. Earlier Investigations Some preliminary work in analyzing and measuring spellers was done by hooK and O’Shea. Their study is well summed up in the following quotations: ’’Let us first examine the word lists of the three spellers previously referred to, and see in how far the ordinary text meets the requirements in respect to a spelling vocabulary. The relation of the vocabularies of the three texts, excluding proper names and foreign terms, is as follows: Speller A Speller B Speller Words found in one text only- — . — -= 5785 1082 Words found in Spellers A and B — 633 633 Words found in Spellers A and C — — 132 132 Words found in Spellers B and C— - 3101 3101 Words found in all three texts—— 1613 1613 1613 Total for each text 2521 11132 5928 Total of different words for all texts, 12489 "The variation among these books with respect to extent of spelling vocabulary indicates that the makers of spelling-books differ in their opinions. .....Not only in regard to total number of words are there striking differences, but also in regard to community of vocabulary. Though Speller A has less than forty-three percent of the vocabulary of C, hardly seventy percent of the list is the same as C’s. That is, the chances of finding any given word of A' s vocab- ulary in C are less than seven in ten. Conversely, the chance of finding any given word of C's vocabulary in A is less than three in ten. Of the total of twelve thousand, four hundred eighty-nine different words, only sixteen hundred thirteen, or less than thirteen percent, are common to all the spellers. "If the lists of the spelling-books be tested by the lists secured from the correspondents, it develops that four thousand, three hundred fifty-one different words, or only thirty-five percent of the total ever apoeared in the letters. The results of the investigations of Cook and O’Shea appeared in 1914, thus antedating the copyrights of the spellers used in this project. Perhaps one is inclined to believe that spellers prepared since that date or revised 'Cook, W. A. and O’Shea, M. V. The child and his spelling . Indianapolis: Bobbs-Merrill Company, 1914. Pp. 225-26. I . V . . I . : . ' - 8 - I since then have met all the important negative criticisms made by Cook and O'Shea. Indeed an examination of the prefaces of the ten spellers used in this project might lead the reader to feel that all that is desired in a good spelling book has been incorporated in each one of them. The following quotations bear espe- cially on the problems attacked in this dissertation; Book I. — "The selection of these three thousand words represents a vast amount of labor extending over a period of several years. Thousands of children's compositions from the H-— School and other schools have been examined, and use has been made of the scientific studies of adults' and children's vocabularies by Jones, Ayres, Pryor, Eldridge, Smith, Wool folk, Cook and O'Shea, and Chancel- lor." Book II.— "Free use also has been made of the Ayres list, Dr. Jones list of so-called 'one hundred demons', and other well-known collections." Book III. --"The words finally selected are those that have been found by recognized authorities to be the most useful in everyday life." Book IV. --"The vocabulary of this book has resulted from a comparative study of the lists published by these investigators (scientific) and other lists from reliable sources. The effort has been to include all words needed by the average person, and to review them so frequently that they will be thoroughly learned." Book V. --"Standard word-lists in harmony with Jones, Pryor, Ayres, and other lists." Book VI •—"The words have not been chosen haphazard, but have been selected on the basis of recent experimental studies, showing the commonly used words in the writing vocabulary of pupils and the corresponding spelling-diffi- culty of these words." Book VII. — "In its (the book's) preparation a careful comparison was made of the vocabularies of several of the most popular spelling books of the day in respect to both gradation and selection. Paralleling this, the various recent tests and investigations, notably those of Ayres, Jones, and Cook and O'Shea have been checked." Book VIII.— "The vocabulary of this book is based upon the lists pub- l lished by various scientific investigators in recent years, supplemented by lists from other reliable sources." Book IX.— "The aim throughout has been to present words related to the needs of each grade as determined by the experience and observation of skilled teachers, and by the results of recent scientific investigations as to the words in general use both in and out of the schoolroom." Book X.--"The vocabulary of the lessons is taken from a compilation which Doctor H. has made of nine scientific investigations of the words used in l writing letters. "The words have been graded in the following manner; On the basis of l - 9 - Doctor H.'s compilation of correspondence vocabularies, all of the 4576 words were ranked according to the frequency with which they are used in correspondence On the basis of Doctor A. ' s study of the difficulty of these words in the various grades, the words were arranged in order of ease of spelling. With these two sources of data, the lessons are arranged so that in general the easiest words and those most commonly and frequently used are placed in the lower grades It is clear that the lessons increase gradually in difficulty in each successive grade . " However, an analysis and measurement conducted not long since on five spelling books of recent date of publication raises sharply the question whether the claims made in the prefaces of spelling texts are altogether lived up to in the preparation of the texts. The results of this investigation are summarized in the following quotation: "On the whole this study shows a great range in the size of the vocab- ulary— from 2991 in the Peirce speller to 6270 in the Champion speller. It also shows the significant fact that little relation exists between the mere size of the vocabulary and the number of the most commonly used words included. For example, the Champion speller with 6270 words contains but 49*5 percent of the Anderson list while the Jones Speller with 4532 words contains 6l.3 percent of the list— which fact signifies that the selection of a spelling vocabulary is something more than merely collecting so many words. The study also shows that the ratio of the number of commonly used words to the total number of words in the book varies from 41.8 percent in the Jones speller to 22.7 percent in the Champion speller. In cold facts these figures mean, if we may safely argue from the basis of the Anderson list, that from $Q to 77 percent of the time devoted to spelling within the public school is spent upon words not commonly used " "Such investigations require time and energy; but it can be safely, predicted that until they are made we shall continue to select books which cause children to spend their time upon words which they will probably never use, and we shall continue to select books incorporating inferior methods."® Another investigation made recently is that of Tidyman. He says: The author lias made a study of thirteen spellers now in general use to determine what percent of the words of these spellers are contained in (a) 3324 words common to two or more of the six investigations reported above, and ( ) all the 6250 different words found in the several investigations. The com- parisons were based upon random selections of pages in the texts in two grades, IV and VII. The results show that on the average only 19 percent of the words in Grade VII are contained in the smaller list, and only 35 percent are con- tained in the larger list. For Grade IV the corresponding figures are 57 percent and 74 percent. That is to say, about four-fifths of the words of Grade VII and wo-fifths of the words of Grade IV are not found in the list of 3324 words com- mon to two or more of the six investigations; about two-thirds of the words of Grade VII and one- fourth of the words of Grade IV are not found even once in the 6250 words of the six different investigations."' Jfaody, C. Application of scientific method in evaluating the subject matter of spellers," Journal of Educational Research , 1:127-28, February, 1920. 7 F * The teaching of spelling. Yonkers-on-Hudson: World Book Com- pany, lyly . p. y. ■ . - 10 - On the one hand, such studies as the preceding ones show how investi- gators were attempting to arrive at a technique of analysis of spellers and an adequate means of evaluating the content of the books after it had been discov- ered, while textbook makers grasping at the results of the investigators were claiming to have exemplified these results in their textbooks. On the other hand, the school people were feeling the urge toward something more fundamental and accurate in judging the content and the methods of teaching in the spellers than had been developed. One form in which this feeling expressed itself was in the derivation of score cards for guidance in judging the textbooks. Score Cards The score cards developed by the school people served, and still serve, in the absence of something better, an excellent purpose. They are decidedly better than no plan for selecting spellers. However, their chief virtue lies in the fact that they point out the more important items that should be taken into consideration in judging the content and the methods of teaching in the spellers that are being examined. The two score cards that follow are typical of the better ones. Joliet Spelling Score Card Rate I on the basis of 30 II on the basis of 40 III on the basis of 20 IV on the basis of 10. 8 I. Vocabulary— — 1. Number of words. 2. Selection of words. a. Child’s vocabulary. b. Provision for enlargement of vocabulary. c. Necessary vocabulary for future as determined by modern surveys. .30 II. Organization-— 1. Gradation. 2. Grouping. • 40 Stoops, R. 0. "The use of score cards for judging textbooks," American School Board Journal , March, 1918. pp. 21-22. i CJ O') - 11 - a. Sound, syllabication, derivation. b. Use, association, related meaning. c. Unusual spellings, homonyms, syno- nyms, antonyms. d. Number of new words and methods of introducing them. 3. Dictation exercises. a. Word building. b. Application of meaning. c. Attention to visual emphasis. d. Wholesome in thought. 4. Comprehensive and frequent reviews and contests. 5* Appeal to child through a. Illustration. b. Quotations. c. Mechanical devices. d. Interesting exercises. III. Aids in developing independent spelling ability — — — 20 1. Phonics, syllabication, accent, diacritical marking, in so far as they lead to efficiency in spelling. 2. Definite help leading to making of rules. 3. Directions for children and for teachers. 4. Training in the use of a dictionary. 5 • Attention to pronunciation. IV. Mechanical make-up — — — — — — .--IQ 1. Paper and binding. 2. Type and arrangement of page. 3. Number of volumes. 4. Cost of set. TOTAL RANK 9 Cincinnati Spelling Score Card Points to be considered in judging the quality of texts in spelling. 1. The words listed should be those which investigations have shown pupils will need most in their written work at school and after they leave school. . Provision should be made for frequent review of words commonly misspelled. . Words should be listed so as to economize efforts in teaching, i. e., grouped as roots, prefixes, suffixes, etc. 4. Suggestions to teachers should offer means of presenting words with a strong initial appeal, for analyzing difficulties, for discovering types of errors, and for following up spelling difficulties in all written work. o ' | ■ ' ' ■ — ■ ' — Mead, Cyrus D. "The best method of selecting textbooks," Educational Admini - stratio n and Supervision . 4:68, February, 1918. - 12 - 5 * Diacritical marks should be used sufficiently to make children self-helpful in consulting the dictionary. 6. Dictation exercises should receive due attention. After a careful study of these score cards, one concludes that they are merely suggestions of what to look for in a good speller. For example, the Jol- iet Spelling Score Card indicates under the caption "vocabulary" that the number of words and the selection of words should be considered, but nothing is said about what should be the number of words or what words are in the child's vocab- ulary, ho v/ enlargement of the vocabulary should be provided for or what are the necessary words for the future as determined by modern investigation. In this score card "gradation” is to be considered but how it is to be determined is not indicated. Looking at the captions of the Cincinnati Spelling Score Card, one sees that "The words listed should be those which investigations have shown pupils will need most in their written work at school and after they leave school." Granting that we have by investigation discovered the words that pupils should study, we have the whole problem of analysis of spellers to find out how these life words overlap the speller words. This dissertation attempts to fill the breach and to supply some of the needed information in this respect. Again, this score card says that "dictation exercises should receive due attention.” This assumes two things: (l) that the judge knows how much attention should be given to dictation, and (2) that he knows how much is actually given in the speller that is being judged. In cur project we make no attempt to answer the first question, but we devote particular attention to the second. Similar criticisms could be made of many of the captions in the two score cards, but it is our purpose merely to call attention to some of the prob- lems that must be solved before the score cards can be used most efficiently. In other v/ords, the score cards deal with the qualitative aspects of the spellers? while our project begins with the qualitative aspects as a point of departure and finds quantitative expressions for some of the more important aspects of - 13 - spoiling textbooks for the elementary grades. The earlier quantitative investigations of spellers were limited in general to one phase of the question, namely, the selection of the words in the texts. Moreover, these investigations were for the most part sharply limited either in the number of texts or in the number of school grades considered. The important questions of the gradation of words and of the pedagogy of each book were left untouched. As to the lack of consideration of the gradation of words, the failure was due in part to the fact that proper instruments for measuring the difficulty of words had not been derived at the time the investigations were made. Nevertheless, the work of these early investigators and the efforts of the school people to derive a satisfactory score card for measuring the texts in spelling should not be underestimated. They blazed the way for future inves- tigators by pointing out some of the important phases of spelling books that should be investigated. Further, they gave clews as to the proper technique to be used, and the nature of the instruments of investigation that should be devel- oped. They showed that we should have measures for the selection of the words, for the gradation of the words and for the pedagogy that is exemplified in the texts. Bases o_f the pro ject The bases or points of departure of this project are the investiga- tions (a) in the selection of words, (b) in the difficulty of words, and (c) in the spelling ability of the children in the different grades. These three lines of research give a body of valuable data and furnish standards by which it is possible to arrive at somewhat more definite conclusions concerning spellers, than concerning any other kind of textbook. The selection of words . ---In the selection of words that should con- stitute the spelling vocabulary for the elementary schools, there has been a - 14 - long line of detailed researches. Among the more careful ones should be men- tioned Ayres's 10 work in which he made up a list of one thousand words occurring most frequently in the written correspondence that he examined, and the words derived by other investigators from similar sources. Cook and O'Shea 11 published lists aggregating 3200 words derived from an extended investigation of the written correspondence of thirteen adults. The frequency of occurrence per two hundred thousand running words was published with the words. This made it possible for other investigators to make extended use of the data. 12 Eldridge published six thousand common English words which he derived from an examination of newspaper articles. He, too, gave the frequency of his words, thus making it possible to compare his work with that of the investigators mentioned in the preceding paragraphs. Under the direction of Professors Briggs and Kelley of Teachers College, 13 the" second and third thousand most frequently used words" were determined by compounding list 5 of the Eldridge list and lists 1, 2, and 3 prepared by Cook and 0 Shea. From the list obtained in this manner, Ayres's thousand words, Jones’s "demons" and certain other words were excluded. 14 Anderson derived three thousand words from an exhaustive investiga- tion of the correspondence of adults. He gave the frequencies of his words. Ayres, L. P. Measuring scale for ability in spe lling. New York; Russell Sage Foundation, 1915* Cook, 'J7. A. and O'Shea, M. V. The child and hi s spelling . Indianapolis: Bobbs-Merrill, 1914. 12 ^ Eldridge, R. C. Six thousand common English words . (Published privately.) Niagara Falls, 1911. ' 13 "Sixteen spelling scales," Teachers College Record . 21:337-91, September, 1920.' 14 Anderson, W. N. The determination of a spelling vocabulary based upon written corres pondence . (Unpublished doctors' dissertation.) University of Iowa, 1917* f - 15 - By combining two or more studies it is often possible to secure a more useful product than was afforded by the component studies. As has been indicated, tnis was done by Ayres in deriving his so-called "thousand commonest words" and again by the Teachers College students in deriving their so-called "second and third thousand most frequently used words." It is clear that we may make one more synthesis— one that is obviously intended by the authors of the "Sixteen Spelling Scales" — and assuming that these words are what they are claimed to be in the reports, we may obtain the three thousand most frequently used words. This may be done by combining the Ayres "thousand commonest words" and the Teachers College "second and third thousand most frequently used words." The frequencies of these three thousand words have a working reliability expressed quantitatively and we can assume that they form a measure of the words that should be taught to the children in the elementary schools. The frequency of these words was derived from their use in their natural setting, that is, in correspondence, in newspapers, in the Bible, and in other literary sources. We thus have a quantitative expression of the demands of society concerning the use of these words. The Difficulty of Words We must determine the suitability of the words in the spellers not only by the demands that society makes upon the schools, but also with reference to the ability of the pupils. In general, with advance in grades the pupils are able to learn to spell more difficult words. Therefore, we should have a measure of uhe difficulty of the words in each grade in order that we may determine whether or not they become more difficult with advancing grades, and whether or not they are in this respect adapted to the ability of the pupils. However, even though the difficulty of the words increases with advancing grades, it does not follow that they are best suited to the ability of the pupils, because the words in their entirety may be too easy, or too difficult. That is, we must have a - 16 - measure of the difficulty of words and also a measure for the spelling ability of pupils; and these two measures must be expressed in the same units. Fortu- nately for our project, a considerable body of data on the difficulty of words and on the ability of pupils has been derived by careful researches. We shall cite the investigations dealing with the ability of pupils in the following sec- tion and in this section we shall cite only some of the more important investi- gations of the difficulty of words. Buckingham 1 ^ determined by extended experimentation the difficulty of a small, selected list of words and expressed their difficulty in terms of dis- tances from an estimated zero of spelling difficulty. Ayres determined the difficulty of his thousand words and published a spelling scale. For the "second and third thousand most frequently used words" in the Teachers College Record 1 "^ the spelling difficulty of each word was determined and expressed in terms of distances from the Buckingham zero of difficulty. Ashbaugh took the three thousand words derived by Anderson and found by extended work the difficulty of each one for each grade. We decided to use as our standard for measuring the gradation of the words in the texts a combination of the Ayres "thousand commonest words" and of the Teachers College "second and third thousand most frequently used words." The difficulty of the words in the Teachers College list was expressed in terms of units of distance from the Buckingham zero, but Ayres expresses the difficulty of his words in a different way. It was possible, however, by the aid of a table 15 ~~ ~ — — — — — ■ ~ — --- * — " ■ ■ ■ — Buckingham, B. R. Spelling ability; its determination and distribution. Teachers College, 1913. 16 Ayres, L. P. Measuring scale for ability in spelling . New York: Russell Sa°-e Foundation, 1915 . 17 "Sixteen spelling scales," Teachers College Record , 21:337-91, September, 1920. 18 Ashbaugh, E. J. " Iowa Spelling Scales ," University of Iowa Extension Bulletin, University of Iowa, 1919. - 17 of equivalents derived by the authors of the Teachers College list to express the Ayres words in the sarne units of difficulty as those used by the Teachers College students. Thus these two lists combined give, as has been said, a three- thousand-word list in which the difficulty of each word is expressed in units of distance from the zero of spelling difficulty. This greatly expedited our in- vestigations; and we therefore chose to make the combination in the manner indi- cated rather than use the Ander son-Ashbaugh list and make the calculations necessary to express the difficulty of all the words in unit distances from zero. Spelling Ability A third line of investigation has added materially to the rigor with 19 which the present project could be carried out. Buckingham ' calculated the median spelling ability of the pupils in each grade and expressed it in terms of distances above an estimated zero of spelling ability. This makes it possi- ble to pass reasonably reliable judgment as to the suitability of the words in a given grade for the pupils of that grade. 2o Ayres in deriving his scale for spelling ability showed increments of ability from a lower grade to a higher one although he did not attempt to locate the zero point or the point where spelling ability begins. 21 Ashbaugh xn deriving his spelling scale showed that spelling ability increased from grade to grade, but he followed Ayres and did not attempt to locate the zero point. These investigations of the selection of words, of the difficulty of words, and of the grade ability of the pupils in spelling offer unique advan- tages to one carrying on a project in analyzing and measuring textbooks in spell- ing that are not found at present in any of the other textbooks in use in the elementary schools. Realizing these advantages, we have undertaken the present ^Buckingham, Ibid. 20 Ayres, Ibid. 21 “'Ashbaugh, Ibid. - 19 - CHAPTER II PROCEDURE IN ANALYZING THE SPELLERS Sp ellers Analyzed In planning our project we finally decided after considerable pre- liminary work to analyze and measure within limits heretofore specified ten recent, popular spellers designed for the grades in the elementary schools of the United States, The clearest principle of selection was that the texts should be of recent publication. As far a3 popularity wa3 concerned we depended on the statements of publishers. We made no effort to determine the amount of sales of the books. Therefore, the fact that we selected a given book does not mean that it had a larger volume of sales than some other book that we did not select. All of the ten books except two were published by different companies, A conspectus of the ten spellers used in this study is shown in Table I. Procedure in_ Analyzing the Spellers A trial set of instructions to clerks for copying off the words of the spellers was drawn up. After experimenting with this set on a number of pages selected at random from three or four of the spellers, the final instruct- ions were written. Questions of interpretation that arose as «he &ork pro= greased were noted for each book. Instructions for copying contents of spellers on cards. --The final set of instructions follow: 1. All words, abbreviations, and contractions presented in the spellers in columns, in dictation, in quotations, and in phrases when clearly set out as lessons for purposes of teaching the spelling of the word3 con- tained therein, were copied on inch by 3 inch cards provided for the purpose. The usual method of indicating a lesson unit is by placing ' . \ ' . ' ; . TABLE I SPELLING TEXTBOOKS STUDIED Code Number Text Author Date of Copyright Publisher Number of Books For Grades Number of Pages I Essentials of Spelling Pearson and Suzzallo 1919 American 1 2-8 xii + 196 II The New-Idea Speller Leiper, Foster, and Weathers 1919 Ginn 1 2-8 xii + 236 III The Ideal Speller (Revised) Willard, Richards, and Brophy 1919 Barnes 2 2-8 xvi + 469 IV Everyday Words Wohlfarth, Pratt, and Wetherow 1916 World 1 2-8 208 V A New Spelling Book Alexander 1918, 1919 Longmans 1 3-8 xxiv + 168 VI £1 son’s Spelling Book El son and Haste 1917 Berry 1 3-8 ix + 282 VII Aldine Speller 3- Bryce and Sherman 1916 Newson 4 1-8 304 VIII New-World Speller 13 Wohlfarth and Rogers 1908, 1910, 1917 World 3 1-8 344 IX Elementary-School Speller Hunt 1916 American 1 3-8 176 X Horn-Ashbaugh Spelling Book Horn and Ashbaugh 1920 Lippincott 1 1-8 105 Material for Grade 1 omitted from consideration. Material for Grade 1 and for grades beyond the eighth omitted from consideration. £> ■ ' i! - pl- over a group of spelling material "Lesson 4" or "2'’ in bold face type, or "Review Lesson" or "Supplementary Words", etc, 2. The words in "Memory Gems" such as Excelsior by Longfellow, Old Ironsides , by Holmes, and in long quotations such as ten lines from "Here is the nation God has builded by our hands. What shall we do with it!--- — " by Woodrow Wilson were omitted. In such cases there was no suggestion that the mater- ial wa3 to be used for spelling purposes. 3« Ayres Spelling Scale when printed as such was omitted. In these cases the scale is evidently presented for purposes of reference and measurement. The words of the scale are presumably presented elsewhere for teaching pur- poses, 4. The following were omitted when copying the words occurring in dictation exercises, quotations, letters, etc. (but not if they appeared in column words): a, an, and, are, at, be, can, day, do, go, he, in, is, it, me, my, of, on, she, the, up, us, we, you. These 24 words constitute over 27 percent of the total running words. They offer almost no difficulty in spelling above Grade 2 as shown by the Ayres Spelling Scale and the Iowa Spelling Scales, 5. Words headed Vo cabulary were omitted because words thus headed are intended undoubtedly by the authors as a "finding list" to indicate what words are presented in the grade or book rather than for lesson purposes. Items indicated on the cards: — In order that definite and extended in- formation about each word in the spellers might be obtained, a set of symbols was devised for the clerks to enter on the cards at the time each word was written. These symbols included the code number of the book wherein the word was found, the grade for which it was offered and the form of presentation as in columns, in dictation, etc. In the upper left-hand corner the code number of the speller was writ” ten in Roman numerals. The grade for which the word was intended was written in Arabic numerals to the right and at the same height as the word. The letter or letters indicating the form of presentation in which the word appears was written below the number indicating the grade. One of the cards is reproduced below. c l A* < < * f C ( < ( ( ( t f t ( < i j i 1 1 . _ r>o „ I receive 6 C 3 " No difficulties were experienced in making entries for the book and the grade in which the word was found. However, this was not the case in making proper entries for the form of presentation. For some words it was a question of judging whether they should be classed as dictation or dictation review ; for others, whether they should be classed as word analysis , word building , or plural formation; etc. Consequently, we are assuming the risk of being tedious in the explanations of the categories in which were placed the different forms of presentation. In the first place we found that it would require an extended and complicated set of symbols to indicate all the forms of presentation in one book and that the series would have to be much more extensive and complicated to distinguish between all the different forms of presentation in all the ten spellers used in this study. Furthermore, lack of use by the authors of a com- mon terminology in designating the different forms of presentation made the task still more difficult. Frequently the form of presentation had to be assum- ed from a general consideration of what had gone before and what followed a given group of words, After a general survey of the forms of presentation in the ten spell- ers, it was decided to make entries as follows: - 23 - a. For all the words presented in column form and clearly for none of the purposes indicated by other symbols, the clerks were to enter C/Column) on the cards. b. For all words in column form clearly indicated for review purposes by such headings as "Review", "Spelling Match", etc., the clerks were to enter CR( Column Review) on the cards. c. For all words in column form that were clearly indicated as additional words to give the children if more work was needed for the year because the preceding regular lessons had been completed, the clerks were to enter the symbol £ (Supplementary) on the cards. d. For all words appearing in phrases, quotations and sentences construct- ed around a given group of words, and not clearly indicated for review purposes, the clerks were instructed to enter I) (Dictation) on the cards. 9, For all wordB appearing in dictation but clearly indicated for review purposes by usually being headed "Dictation Review", the clerks were to enter DR (Dictation Review) on the cards. (Note--Signif icant words in dictation (D) and in dictation review (DR) were underlined on the cards whenever it was possible to determine that they were significant, e.g. when they were underlined, or put in columns in connection with dicta- tion in the same lesson). f. The clerks were instructed to enter on the cards the symbol WB^ (Word Building) for all words that are clearly indicated to be for the pur- pose of assisting the pupils to understand and spell them by adding prefixes, suffixes, etc. The devices used in the textbooks for indicat- ing these words are for example: placing them under the heading, "Word- Building" ; or accompanying them by statements such a 3 " Abl e means worthy or able to be. Add able to each of the following words and note the change in meaning. Be careful to make no mistake in spelling the derivatives," g. The clerks were instructed to enter on the cards the symbol WA (Word Analysis) for all words that are clearly indicated to be for the pur- pose of assisting the pupils to grasp their meaning by separating them into roots, prefixes, etc. The devices used in the textbooks for in- dicating these words are for example: placing them under the heading "Word Analysis"; or accompanying them by statements such as,. " Ant and ent means one who; that which. For example, resident means one who resides. Analyze each of the following words so as to show from what verb it is derived," h. For homonyms grouped together in columns and so indicated in the text by the device, "Homophones defined" or by similar devices, the clerks were instructed to enter on the cards the symbol Cli (Column homonyms) ; for all homonyms appearing together in dictation and clearly indicated as such by the context or by the heading, the clerks were to enter on the cards the symbol DH (Dictation Homonyms); and for such words sim- iliarly indicated and with the additional notation that they are for review purposes the symbol RF[ (Review Homonyms) was used regardless of whether in column or dictation form. - 24 i. The clerks were instructed to enter the symbol PF (Plural Formation) on the cards for words clearly indicated to be for the purpose of teach- ing the spelling of their plural forms. The devices used in the text- books for indicating these words are for example: placing them under the heading "Plural Formation"; or accompanying them by statements such as, "Following the rule, add es^ to these words, and use the new forms in written sentences of your own." j. For all words clearly indicated by such devices as the heading "Phonic Series", or by groups of words such as head, bead , rot , trot , tie , hie , lie , etc, the clerks were instructed to enter on the cards the symbol PH (Phonics). Alphabetizing and making mast er c ards . — V/hen all the words in a spell- er had been written on cards, they were arranged in strictly alphabetical order for each grade. Thus all duplicates were automatically brought together. For words appearing more than once in a particular grade for a given book, all the data pertaining to the word were transferred to one of the cards on which the word already appeared. This card was called the master card. A sample of a master card is shown below. master card I 2 girl 1C 2D ' 1DR This master^ card reads: In Book I Grade 2 the word girl appeared once in column form, twice in dictation form and once in dictation review form. These master cards were left in the filing trays, while the dupli- cates were taken out, sorted thru once to find any master cards that might have been placed among them through error, and then discarded as far as this project is concerned. Eliminations : — » The files of alphabetized master cards were gone through and eliminations were made according to the following directions: — - — — -■ — - - 25 - 1 ! different forms of the same word are dif f eront wo rds except that the simple adding of £ to a noun or verb does not constitute a different form. If a book contains two forms identical except that one has an added & and the other has not, reject the form in s, and enter its data on the master card for the simple form without s. "if a book contains either the simple form alone or the form in £ alone, retain the word as it stands , . If a book contains three' or mTre forms of the same word, one of which is the simple form and one of which is the form in s, re- ject the form in s_ and enter its data on the master card for the simple form without s_. If it contains two or more forms none of which is in £, retain them all. Exception; Include both the singular and plural of nouns whose sing- ular ends in £ preceded by a vowel, cr in o if a spelling difficulty might arise when the plural is formed, e.g. chimn ey and chimneys; solo and solos. 1 1 2. Take out all proper names except days of the week, names of the months, and Christmas, Easter, Thanksgiving, and Hallowe'en, 3» Take out all compound words both parts of which are elsewhere listed in simple forms in the same book unless a change of spelling or pronuncia- tion occurs in the compound. If the parts are not listed in simple form, separate them and write on cards. 4. If the same word appears two or more times and one of the appearances is capitalized, take out the card for the one that is capitalized and enter its data on the master card. 5» Retain words spelled with an apostrophe. 6. Exclude foreign words and phrases. The number of different words in each grade was determined by count- ing the master cards. The number of column occurrences was found by counting all the on the master cards; the number of dictation occurrences was found by counting all the P's , etc. Later the cards, as alphabetized by grades in a given book, were thrown together in strictly alphabetical order for the book as a whole, and various countings were made. Other details of procedure will be discussed later in connection with certain tables. - 26 - I CHAPTER III EXTENT OF THE VOCABULARY One of the important problems connected with spelling texts is the ex- tent or size of the vocabulary to be- mastered by the children. It indicates the textbook maker’s idea as to the number of words that society requires to be taught to children; his concept as to the ability cf the pupils to master a given number of words during their school career; and in part the influence that mod- ern research in word selection has had upon him in the construction of his spell- er. The extent of the vocabularies found in the texts used in this project gives an opportunity to compare our findings with those of ether investigators who have covered the same field. Extent of the Vocabulary in Each Book . In our analysis of spellers a careful count of the different words in each book was made after the words had been arranged in alphabetical order for the bock as a whole. By this method no word was counted twice or mere and it is possible to report within a reasonable degree of accuracy the extent of the vocabulary in each book. Table IX shov. ; s the number of different words in each book . - 27 TABLE II NUMBER OF DIFFERENT WORDS IN EACH BOOK Book Number of Dif f erent words Book Number of Di f f erent words ! i 3968 VI 5328 ii 5001 VII 4545 hi 478? VIII 4767 j IV 4742 IX 5915 j * V 5911 X 4159 1 , 1 Table r eads: In Book I there are 3968 different words; in Book II there are 5C01 different words; etc. The number of different words ranges from 3968 in Book I to 5915 in Book IX. This is a wide range of approximately 2,000 words and it should be stated that Book I is designed for Grades 2 to 8 inclusive while Book IX is de- signed for only Grades 3 to 8 inclusive. The average size of vocabulary is 4900, the median 3ize is approximately 4750, and the modal number of words in the vocabulary of a speller is approximately 4750. These data indicate that the makers of the spellers are at variance as to the number of words that the child needs to meet the demands of society and as to the ability of the pupils to learn efficiently the spelling of words. We have some data concerning the number of words that should be taught. Ayres^ on the basis of his researches "decided to limit the foundation vocabu- lary to 1,0C0 words." Anderson^ limited his vocabulary to 3105 words, The Teachers College List^ was limited to two thousand words beyond the Ayres 1,0CC 1 Ayres, L.P. Measurement of ability in spelling, p.9, 2 j..- h - e . t fterrcination of a spelling vocabulary base d upon written correspondence. ^Unpublished doctor's dissertation) University of Iowa, 1917. 3 „ Sixteen spelling scales." Teachers Collc-e Record . September, 1920. - 28 - 4 word. Jones found that the largest vocabulary of the eighth grade student was 2812 words. Tidyman^ says: "Thus it appears that a writing vocabulary of 4,000 or 5,000 words is adequate for the most exacting and varied demands that ere likely to be made upon the average child, and that the thorough teaching of 3,000 to 3,500 carefully selected v/ords is about all that should be expected of the ele- mentary' school, provided that in addition the child is taught how and when to use the dictionary Horn- on the basis of compilations made from eight vocabulary studies placed the limit at 4C 52 and said that they "should be used as the basis of spelling instruction in the elementary school until more elaborate investiga- tions are made." Therefore it seems that a 4,000 v, r ord vocabulary would be the upper limit and a 3,000 word vocabulary the lower limit on the basis of the social de- mand as suggested by these investigations. If this suggestion is reasonable, only two of the texts examined by us approach closely the standard. These two are Books I and X, and they lie approximately at the upper limit of cur suggested standard size of vocabulary. As a result of the studies of the memory span of children the tendency in^recent years has been to teach a small number of new words a day. • If we assume that two new words a day are taught in Grade 2 and 3, three new words a day in Grades 4 and 5 and four new words a day in Grades 6, 7 and 8, and that the school year is l 80 days we shall Have taught 35&C different words to the child who completes Grade 8. These data based on the assumed ability of the pupil to learn new words bring us to practically the same conclusion as did the considera- 4 Concrete investigation of the material of English spelling. 5 The teachir. ; of spelling , p. °. 6 The principles of method in spelling as derived from scientific investigation. The Ei ghteenth Yearbook of the National Society for the Study of Education, Pt. II, p. 57. n - Horr,E: Ibid , p. 58 . - 29 - tions based on the danands of society, that is, that 4,0C0 words set the upper limit to the extent of the vocabulary . Therefore, from the sociological basis and from the psychological basis, we conclude that Books I and X most nearly con- form to the proper size of the vocabulary. The other books on our list of ten have vocabularies larger than 4,5C0 words. We conclude, on the basis of the hypothesis that 4,0C0 words is the upper limit, that they are attempting to teach too many words. It is possible to compare the size of the vocabularies in cur ten spell- O ers with the size of the vocabularies in a few other spellers. Cook and O'Shea"' found that their Speller A had a vocabulary of 2,521 words, Speller B had a voc- abulary of 11,132 words, and Speller C had a vocabulary of 5*928 words. The vocabulary for Speller A was smaller than the vocabulary for any of our ten spell- ers, the vocabulary for Speller C was practically the same size as the vocabular- ies in our Books V and IX, while their Speller B had a vocabulary about twice as large as either of our two largest— Books V and IX. However, the data derived by Cook and O'Shea' are not strictly compar- able with our data, because of a difference in the methods of selecting the vocabularies. It would seem, indeed, that Cook and O'Shea are not consistent v/ith themselves. Their Speller A contained only column words, hence no problem in selection presented itself. But in their speller B only column words were used regardless of the fact that half the significant words we re presented not only in column but also in adjacent sentences. In their Speller C, however, which contained a large number of dictation and completion exercises, "everything save directions to pupils" was used. In our investigation, however, we used all the different words whether or not they w r ere in column or dictation form of pre- sentation. If their spellers were representative of those in current use in 1913 8 Cook, W.A. and O'Shea, M.V. 9 Ibid^, pp. 148-9 The child and his spelling , p . 22 5* - 30 - and 1914, and their data comparable with ours, and if our ten are representative of the present practice, then it follows that there was more variation in the size of the vocabularies among the spellers at that time than exists now. The investigations made by Woody 10 were more recent than those of Cook and O'Shea. His results are as follows: TABLE III NUMBER OF DIFFERENT WORDS IN EACH BOOK ( WOODY) n”- Sp ellers Grades Total words in Book Champion 3-8 5*872 Peirce 3-8 2,862 Merrill 3-8 5,550 Aldine 1-8 4,436- Aldine 3-8 3 » 4C0 Jones 2-3 4,532 / j While these data do not show as wide range in the size of vocabulary as did the spellers examined by Cook and O'Shea, there is nevertheless a wide range from 2,362 to 5*872. That is, the largest book has twice as many words as the smallest book. This range is wider than ours which as has been said is from 3 i 968 to 5 i 915. However, Woody's data and curs may not be strictly comparable because the sources of the words may not be exactly the same. We used all the different words whether in column or in dictation form, while we are not able to determine from his description of the sources of his words whether or not he did the same. One of the books examined by Woody and one of our ten is the same book^. the Aldine. We found 4,545 different words for Grades 2 to 8 while Woody es- timated the number for Grades 1 to 8 to be 4,436 and for Grades 3 to 8 to be 3 * 4C0. It would be difficult to determine on the basis of Woody's data what the Woody, C. Application of scientific method in evaluating the subject matter of spellers." Journal of Educational Research . 1:119-128, February, 1920. - 31 number of different words for Grades 2 to 8 inclusive would be, but we assume that it would be between 3>4C0 and 4,436 words, which would be appreciably less than the 4,5^5 different words that we found for Grades 2 to 8. This discrepancy may be due to slightly different sources for the words. We used the words found in both column and dictation forms of presentation, while Woody may have used the words found in column only. However, from his descrip- tion of the sources of his speller words, we are not able to determine whether or not he used dictation material. If he did not, then this is perhaps the explana- tion of the discrepancy. Extent of^ the Vocabulary in Each Grade An examination of the extent of the vocabulary or the number of differ- ent words in a grade reveals (1) the teaching and learning load that confront the teacher and the pupils; and the extent to which the authors have used a common standard for the number of different words that should be taught and re- viewed in each grade. The reader may be misled if he does not note carefully that we used the term "taught and reviewed" which applies to words in all the grades except Grade 2 in the spellers having Grades 2 to 3 and to words in all grades except Grade 3 in the spellers having Grades 3 to 8. A determination of the number of words to be "taught"— i. e. , the number of new words— for al] the grades of ten spellers would have entailed a very large additional outlay in time and money. The problem, however, is straight forward and requires only clerical service nec- essary to derive the data. No doubt some investigator will undertake the task. Ne suggest that makers of spellers use some common device to indicate in each grade the words that are presented there for the first time in the book. The word might, for example, be underlined or printed in bold face. With the preceding cautions and limitations in mind we present the number of different words in each grade in Table IV. c , . ' • V V 1 J . . X, . ‘ . .. e • C . . i . 1 'i, > i. ' • : * ' - u» . . X . *t ' 32 - TABLE IV NUMBER OF DIFFERENT WORDS IN EACH GRADE Grade Book I II III IV V VI VII VIII IX X 2 377 701 370 420 (a) (a) 576 569 (a) 419 3 589 866 750 930 951 1024 822 1092 1030 662 4 868 950 1307 1229 1235 1392 993 1379 1166 762 5 965 926 1459 1256 1589 131? 1117 1415 1100 751 6 1032 1057 1508 1294 1864 1236 1142 1483 1367 766 7 794 1139 659 858 I960 1367 1161 1582 1723 718 8 1240 1413 1345 1067 1950 1395 843 1599 1914 1143 ( a.) 'These books do not provide for spelling in Grade 2. Table reads : In Grade 2, Book I, there are 377 different words; in Book II there are 701 different words; etc. I - 33 - It will be interesting to note the range of the number of different words in each grade through the different books. In Grade 2 the range is from 37C to 7 Cl different words; in Grade 3 from 589 to 1092 ; in Grade 4 from f62 to 1392; in Grade 5 from 751 to 1589; in Grade 6 from 766 to 1864; in Grade 7 from 659 to I 96 O; and in Grade 8 from 843 to 1950* Some books have twice as many- different words for a given grade as other books and for some grades one book has three times as many words as another book. In general there is an increase in the number of different words in advancing grades. However, there is shown by the table two or three exceptions that should be noted. .The number of words decreases in Grade 7 Book I and no explanation is found. For Grade 7 Book III it should be noted that long quota- tions given in the book were not copied on cards in the original work, hence it is to be expected that the number of different words would decrease. This also accounts for the relatively small number of different words in Grade 8 of this book. A series of straightf orward calculations made on the data in the table will reveal the range in the number of different words a day that the child would meet if we allow 180 days for the school year. In Grade 2 the range would be from 2 different words a day in certain books to 4 different words a day in other books; in Grade 3 from 3 to 6 different words a day; in Grade 4 from 4 to 8 words; in Grade 5 from 4 to 9 words; in Grade 6 from 4 to 10 words; in Grade 7 from 4 to 11 words; and in Grade 8 from 5 to 11 words. In Grades 2, 3 » 4, 5 » and 8 one book has twice as many different words a day as another book. In Grade 7 one book has three times as many different words as another book and in Grade 6 one book has two and a half times as many words as another book. Figure x shows the variation in the number of different words in Grades 2 to 8 of typical spellers. - 34 - />? 76 77 /e 75 74 rl 77 1 7 6 4 f/yore / A/v/Tider of f)7ffere/7f/4frrz/s 6y Grades 7/7 7yp/oa/ Spe//efs / Grade 4 ... 1 7 - 35 - Hera we see wide diversity of opinion concerning the number of differ- ent words in a given grade. One speller has, for example, 843 different words in Grade 7 and another speller has i 960 . Surely such wide diversity of practice can not be justified on the basis of experimental data. This suggests the ur- gent need for research work to determine the optimum number of different words that children of a given grade can learn to a reasonable degree of efficiency, and the number of review words that should be presented. - < , . : - 36 - CHAPTER IV SELECTION OF VOCABULARY One of the most popular subjects now before school people and research workers in education is curriculum construction. Elaborate methods and extended formulas have been used widely in the search for scientific methods of curricu- in curriculum construction is found in the efforts of research workers to deter- mine the words that should constitute the spelling vocabulary to be taught to the elementary school children. One of the moot questions is: What should be the source material from which to select the vocabulary to meet the needs of society? Knowles used the Bible and certain literary selections and formed his list from the basis of the most frequently occurring words. Eldridge used newspaper articles and selected 6000 of the most frequently occurring words. Cook and O’Shea based their words on frequency in adult correspondence. Ayres selected his vocabulary on the basis of frequency, in part from adult correspondence and in part from a synthe- sis of the three preceding lists. Anderson derived his list on the basis of frequency in adult correspondence. The Teachers College students derived their words from a synthesis of the Eldridge and the Cook and O'Shea lists excluding the Ayres list of "thousand commonest words," Jones’s "demons" and certain other words. Consequently, the frequency of use in adult correspondence and in news- papers was the basis for the Teachers College lists. Jones obtained his ^The” Fourteenth, Sixteenth, and Seventeenth Yearbooks of the National Society for the Study of Education, Part I, (Public School Publishing Company, Bloom- ington, Illinois) and Charters, W. W. "What has thus far been accomplished and is now available for the readjustment of school curricula." proceedings of the High-School Confe r- ence , 1918, University of Illinois, Urbana, Illinois. lum construction 1 An excellent illustration of the use of scientific methods - 37 - vocabulary from the frequency of use in the themes of school children; Woolfolk obtained his from the frequency of occurrence of errors in the written work of elementary school children. All of these investigations gave lists of words as found in their natural setting, i. e., in sentence formation* The sources may be grouped as follows: frequency in adult use — corre- spondence, newspapers and literature; frequency in children’s correspondence — themes and spontaneous compositions; and frequency of error in children’s writ- ten work. It seems clear that none of these sources taken exclusively as a basis for word selection is the best. What is needed is a synthetic vocabulary selected on the basis of frequency of occurrence in children's written work which gives the words the children need to spell in their present social activities; then, to avoid teaching words that are already learned incidentally, the frequen- cy of error should be considered; and, in order to provide for enlargement of vocabulary, the frequency of occurrence of words used in adult correspondence should likewise be considered. For purposes of evaluating the selection of the vocabularies in spell- ers one should know the words the children need to spell and also the frequency and the difficulty of these same words. Such an ideal list with the sources we have indicated does not exist at present. In the absence of such, we decided to use a synthesis of the Ayres "thousand commonest words" and the Teachers Col- lege "second and third thousand most frequently used words" as our measure of the selection of the vocabulary of spelling books. This was done for expediency and because all of the spellers indicated in their prefaces that in the selection of their vocabularies use had been made of many of the scientific investigations of spelling including especially Ayres's, Eldridge's, and Cook and O’Shea's — the last two being the basis of the Teachers College list. Selection o_f Vo eabulary by Books The extent to which use was made of the Ayres and the Teachers College J - 38 - lists is shown in part in Table V. Book I has 994 of the Ayres "thousand commonest words" which is the largest number contained in any of the books. Book II has 808 of the same list and this is the least number found in any of the spellers. As far as the Ayres so-called "thousand commonest words" are a reliable measure of the demands of society upon the teaching of spelling in the elementary schools, we may say that on the whole the texts have met the demand exceedingly well. This is especially true of Books I, VII, and VIII. It is justifiable to omit from the speller from twenty to twenty-five of the words of the Ayres list because they offer compara- tively no spelling difficulty to the children even in the first two grades (that is, Grades 2 and 3) for which we have data on the difficulty of words. There is another possibility that the spellers would reveal no more than 976 words because we excluded from our study, when the words we re being written on the cards, twenty-four of the easiest words in the Ayres list unless these words were found in columns. (For a list of the excluded words the reader should refer to Chap- ter II.) On the basis of the Teachers College words, the texts do not make such a favorable showing. Of course, we are assuming that this list of words is what it purports to be, that is, the second and third thousand most frequently used words in newspapers and adult correspondence. The range in the number of these words is from 1019 in Book II to 1323 in Book VIII. All things considered, it is remarkable that the range is so narrow for the ten books. As one glances down the column showing the number of Teachers College words in each text, he cannot fail to observe the similarity in the numbers of these words common to the spellers. He will note, for example, that all the books have more than one- half of the 2000 words of this list and that none of them has two-thirds of these words. Column 5 of the table shows the percent of the different words in each 39 TABLE V EXTENT TO WHICH THE WORDS OF THE AYRES- TEACHERS COLLEGE LIST ARE FOUND IN THE TEXTBOOKS Book Number of different words Number of Ayres word 8 Number of Teachers College words Number of Ayres- Teachers College words Percent of words in book that are Ayres- Teachers College words a Percent of Ayres- Teachers College words in book b I 3968 994 1052 2046 51.5 68.2 II 5001 808 1019 1827 36.5 60.9 III 4787 957 1C88 2045 42.7 68.2 IV 4742 954 1257 2211 46.7 73.7 V 5911 861 1181 2062 34.9 68.7 VI 5328 917 1103 2020 37.9 67.3 VII 4545 976 1187 2163 47.7 72.1 VIII 4767 983 1323 2306 46.4 76.9 IX 5915 944 1272 2216 37.5 73.8 X 4159 916 1303 2119 50.9 70.6 Entries in Column 4 divided by entries in Column 3 . b Entries in Column 4 divided by 3000. Table reads : In Book I there are 3968 different words consisting in part of 994 Ayres thousand words and 1052 Teachers College words, these totaling 2046 words; the Ayres-Teachers College words form 51.5 percent of all the different words; and 68.2 percent of the Ayres-Teachers College words are common to the text; etc. V - 40 - book that are Ayres-Teachers College words. The range is from 34.9 percent for Book V to 51*5 percent for Book I, and the median percent for the ten books is 42.?. Assuming that the 3000 words of our synthetic list are the words that should be taught first, no matter what others are taught, it follows that 48.5 percent to 65*1 percent of the words now taught are of the kind that should be presented only after the 3000 Ayres-Teachers College words have been attended to. In other words, from 48.5 to 65 . 1 percent (almost l/2 to 2/3) of the pupils' time is devoted to a relatively less important content. It seems clear that something more than l/3 to l/2 of the time ought to be spent in learning a list which is fundamental from almost any point of view. Since there are indications that the vocabulary for a speller may be in the neighborhood of 4000 words, it is clear that devoting anything like l/2 to 2/3 of the time of the pupils to words outside of a fundamental list of 3000 is a faulty arrangement. If we were 100 percent efficient in teaching the words in the Ayres-Teachers College list, it might not be so serious, but we know from results of measuring the children's performances in spelling that the teaching of these words falls far short of perfection. Our data on the percent of the words in the spellers that are Ayres- Teachers College words make a better showing for the spelling texts in regard to their presenting the "most commonly used words" than do similar data published by Woody. He found for the "percent of words in book that are in the Anderson list" the following: Champion speller 22 - 7 ; Peirce speller 26.2; Merrill speller o 30; Aldine speller 41.5; and the Jones speller 41. 8. ^ He concludes thus: "In cold facts these figures mean, if we may safely argue from the basis of the Anderson list, that from 58 to 77 percent of the time devoted to spelling within the public school is spent upon words not commonl y 3 used." Our corresponding figures are 48-5 and 65 . 1 percent. That is, our data 2 ; — — - - Woody, C. "Application of scientific method in evaluating the subject matter of spellers," Journal of Educational Research, 1:121, February, 1920. 3 Woody, Ibid. p. 128. — 1 41 are more favorable to the selection of words found in spellers. However, the figures obtained by us and by Woody may not be closely comparable. In the first place, we do not know what our results would have been if we had used the Ander- son list, although we do know that the two lists purport to be the same thing, namely, the 3000 most frequently used words in adult writing. In the second place, we know nothing as to the basis used by Woody in selecting the words in his spellers. As mentioned before, he may have confined his attention to words printed in columns, but he does not describe his method accurately enough for us to determine. Cook and O'Shea 4 found that "If the lists of the spelling books be tested by the lists secured from the correspondents, it develops that four thou- sand, three hundred fifty-one different words, or only thirty five per cent, of the total, ever appeared in the letters." While their data were derived by a different method of calculation, their measuring list, like ours, was obtained from adult writing and in fact, it composed a part of our measuring list. Their result, however, approaches the results that we found. If the hypothesis holds, that the elementary schools should meet the demands of society in regard to the words to be taught to the children, and if the demands of society are interpreted correctly by the so-called "most frequent- ly usea words” derived by the different investigators, then it may be concluded from the sets of data presented in the preceding paragraphs that the spellers examined so far are not measuring up to what is expected of them. Another view of the way the spellers are meeting the social demands may be obtained by studying Column 6 in Table V. This column shows the percent of the Ayres-Teachers College words that are found in the different spellers. The range is from 60.9 percent of the Ayres-Teachers College words found in Book II to 76.9 percent found in Book VIII. In spite of this range of 16 percent one 4 Cook and O'Shea. The child and his spelling , p. 226. <•- \ • - 42 - feels that the books are not strikingly different in this matter and that on the whole the percent of the Ayres-Teachers College list included in the spellers is rather higher than might be expected. However, if the Ayres-Teachers College list is what it is purported to be, the three thousand most frequently used words, it should be found approximately 100 percent, or entire, in any speller which offers more than 3000 words. Our data again are more favorable to the spellers in general than are the data derived by Woody on the basis of the Anderson "three thousand most fre- quently used words." He showed the percent of the Anderson list included in spellers as follows: Champion speller 45*9; Peirce speller 25.2; Merrill Bpeller 54.0; Aldine speller 63.9; and Jones speller 6l.3." It will be noted that the range obtained by us for our list of "three thousand most frequently used words" is from 60.9 percent to 76.9 percent while the range found by V/oody is from 25.2 to 63. 9 • Our limits on the whole are much higher than his. The data obtained by us, as mentioned before in this section, are not strictly comparable with those obtained by Woody. We considered the words in all forms of presentation — column, column review, dictation, dictation review, etc. While we are inclined to believe that Woody considered only the column form of presentation, he does not make this point clear. If we are correct in assuming that he used column words only, then we are safe in the statement that he did not find all the different words. This statement is based on the fact that when we found the words common to the Ayres-Teachers College list and to the column and significant dictation (underlined, printed in columns in the same lesson, etc.) words in the spellers, we secured an appreciably smaller number than when we considered the words common to the Ayres-Teachers College list and to all forms of presentation in the spellers. Therefore, we conclude that Woody's meth- od of selection if, as is apparently the case, it was based on column words alone, ^Woody , Ibid. p. 121 - 43 - failed to find all the words and consequently that our method is better. However, regaraless of the fact that in many ways our data and his are not strictly comparable, nevertheless they indicate the same general conclusion, namely, that the spellers are not teaching the words that modern research in the selection of the spelling vocabulary indicates should be taught. One of the practical outcomes of the preceding table and discussion is the possibility of ranking the spellers. The schoolman in service is more interested perhaps in the rank of each book than he is in the exact numbers and percents for any given phase of the spellers. ,/e have chosen to rank the spellers on three bases — the percent of words in each book that are Ayres-Teachers College words, the percent of the Ayres-Teachers College words that are found in each book, and a combination of these two sets of data. The rank of each speller in each category is shown in Table VI. TABLE VI RANK OF SPELLERS ON PERCENT OF AYRES-TEACHERS COLLEGE VORDS CONTAINED IN THEIR VOCABULARIES Book (a) Rank' ' on per- cent of Ayres- Teachers College words ( a’i Rank' ' on per- cent of words in book that are Ayres-Teachers College words Combined Rank I 7-5 1 5 II 10 9 10 III 7-5 6 6 IV 3 5 3.5 V 6 10 8.5 VI 9 7 8.5 VII 4 4 3.5 VIII 1 v' 3 1 IX 2 - 8 6 * -1 5 v 2 2 a ^Rank 1 means highest, rank 2 next highest, etc. iabl.e reads ; Book I ranks 7*5 on the percent of the Ayres- Teachers College words that are found in its vocabulary, ranks 1 on the percent of the words in its vocabulary that are found in the Ayres-Teachers College list, and on the combination of the two percents it ranks 5; etc. - 44 - I"t will b© noted that of the two books which mak© the best showing Book VIII has a high ranking in both items and Book X ranks only fairly high on one of the items. Again, it will be seen that the books that rank low do so consistently. This regularity of ranking might suggest to the reader that the measures are dealing with a constant function in the spellers. However, that these functions are independent may be shown by the fact that the size of the vocabulary has practically no correlation with the number of Ayres-Teachers Col- lege words found in the books. The value of r_ derived by using the rank-differ- ence method, is -0.09. Selection o f Vo cabulary by Grade s The preceding study of the selection of the speller vocabularies by books gave a general picture of them as a whole, but we need to go farther and to examine by grades in each book the selection of the vocabulary. With our graded system of schools today, the grade is in reality the teaching unit rather ti.an the whole series of grades. In fact, many textbooks in other fields, as well as in spelling, are not published as single volumes but as series of volumes witn material for one or two grades in each book. Often we find that a school system has a given textbook for certain of the grades and another textbook for certain of the other grades. Presumably, this is because the school authorities believe that one book is especially strong in certain features for a particular grade or grades and that another book is better suited to the other grades. Therefore, it is well to study the texts in spelling by grades. We can determine what use is made of the "most frequently used words" in each grade in each book, how the authors vary in the use of these words, what the general tendency among these practices is, and where the teaching and learning loads are as far as the "most frequently used words" are concerned. In each grade in each book the number of Ayres "thousand commonest words," the number of Teachers College "second and third thousand most frequently ; " , ■ . • .. - 45 - vised words,” and by adding the two preceding, the number of Ayres-Teacher s Col- lege words was determined. It should be noted that ’’number" refers to the dif- ferent words and not to the different words plus the repetitions of these words in the grade. Further, it should be noted that the number of different words in ^ Si ven giade does not mean that these words did not appear in another grade, since a given word was often found in two or more grades — as will be shown in Chapter VI. With these statements and definitions in mind the reader’s attention is called to Table VII which shows the number of Ayres-Teacher s College words per grade. From Table VII we observe that the ranges between the spellers offer- ing the least number of different Ayres-Teacher s College words per grade and the spellers offering the largest number of such words per grade are; Grade Range in number of words Grade Range in number of words 2 174 5 476 3 403 6 542 4 526 7 6l8 8 499 These figures show a wide variation in the number of Ayres-Teacher s College words found in a given grade in the spellers. This is especially true when we consider that the average number of such words per grade does not exceed 600. Since we have little, if any, experimental data on the number of new words that should be taught in a given grade and at the same time the number of words that should be reviewed, we may approach the problem from the angle of current practice. On the supposition that our ten spellers represent the best practice at the present v/e can find the average of the ten for each grade. % - ■ - 46 - TABLE VII NUMBER OF AYRES-TEACHERS COLLEGE WORDS PER GRADE Grade Book I II III IV V VI VII VIII IX X 2 319 493 333 391 (a) (a) 395 404 (a) 370 3 475 524 577 838 738 834 530 733 571 435 4 608 525 792 965 629 703 626 815 565 439 5 780 447 847 893 823 625 680 807 685 417 6 658 464 917 838 818 478 645 768 619 375 7 442 422 224 435 842 457 633 816 682 339 8 368 432 424 423 748 449 281 780 604 559 ( a 1 ) ' ' These books do not provide for spelling in Grade 2. Table reads ; In Grade 2, Book I, there are 319 different Ayres- Teachers College words; in Grade 3, Book I, there are 475 different Ayres-Teachers College words; etc. - 47 - However, this current practice succeeded in teaching only about 2000 of the 3000 words in the Ayres-Teachers College list. In fact, the average number of the words in this list that are presented in the ten spellers is 2102. Therefore, if we wish to present the 3000 words of the Ayres-Teachers College list on the basis of the current practice, v/e shall have to increase the average offerings in each grade. For this purpose we shall have to set up a standard for the in- crease. It seems that the average of 386 words for Grade 2, as shown in Table VIIJ. below, is fairly reasonable, since this would be only a little more than 2 new words a day and since children have learned to spell relatively very few words before they reach Grade 2. The average number and the hypothetical number of Ayres-Teachers College words per grade is shown in Table VIII. TABLE VIII AVERAGE AND HYPOTHETICAL NUMBER OF AYRES- TEACHERS COLLEGE WORDS PER GRADE Ayres-Teachers College words per grade Grade Average number for ten spellers^ 8- ' which teach 2102 Hypothetical' num- ber to teach^) 3000 2 386 386 3 626 888 4 667 946 5 700 993 6 658 934 7 530 752 8 507 719 Total num- ber of oc- currences 4074 5618 ^The averages were calculated from Table VII. (b) The hypothetical number was obtained by; (1) Increasing the occurrences (4074) in the same ratio that the average number of Ayres- - 48 - Teachers College words taught in the spellers (2102) is to the increase necessary in order to teach the 3000 Ayres-Teachers College words. (2) After the occurrences were thus increased to 56l8, the 386 occurrences for Grade 2 were subtracted leaving 5232 occurrences. (3) Then the 386 occurrences for Grade 2 were subtracted , from 4074 occurrences leaving 3688. (4) There- fore, the increase in number of occurrences is 1544 which gives 41.9 percent. (5) Hence, above the second grade each grade-average for the ten spellers was increased by this percent to obtain the hypothetical grade-average num- ber of Ayres-Teachers College words. The sudden drop in the number of words common to the Ayres-Teachers College list in Grades 7 and 8 of Book III is due to the fact that the long quo- tations from literature were omitted. It should be noted for Grade 8 of Book X that the rapid rise in the number of Ayres-Teachers College words is due probably to the fact that in this grade appeared four one-page letters to be used for dic- tation purposes. These facts suggest that the number of Ayres-Teachers College in a grade is influenced by the amount of dictation in that grade. In order to find out the extent of this suggested influence, the correlation between the num- ber of words in dictation and the number of Ayres-Teachers College words found in the grade was calculated on the arrays of the grades in the ten spellers, and it was found that by the Pearson product-moment formula the value of r was +0.48. This coefficient is high enough to indicate that the author of a speller may partly provide for the appearance of these words by including in each grade a liberal supply of dictation material. Of course, the more carefully the dicta- tion material is chosen with respect to the number of Ayres-Teachers College words it contains, the more the spelling of these words will be guaranteed. That the proper selection or composition of this dictation material requires great care, follows without question. Hence, the task of selection should not be placed on the classroom teacher but it should be done by the authors of the spellers. The number of Ayres-Teachers College words found in a grade is - 1 . • . . \ • • . .. > • . . . . ■ .. . . . ■ i ■ •' ; , ■, . . . . . * • ■ ■ ; ... Y ' • . ■ ' ■ . ■ . . ■ j ■ | _ • - • j . .. . ■ . . . . ... . . ' : . .. - 49 - influenced not only by the amount of dictation but also by the number of differ- ent words in the grade. The correlation between the number of Ayres-Teachers College words and the number of different words in a grade calculated by the Pearson product-moment formula on the arrays for the ten spellers is +0.47. This is also a relatively high positive correlation. It suggests that an author by the laws of chance rnay insure a large number of Ayres-Teachers College words by offering a large number of different words in each grade. We have already indicated in this chapter that for these ten books there is no relation between the number of different words in the books and the number of Ayres-Teachers College words. This would seem to be at variance with the statement we have just made. It would appear on the one hand that by in- creasing the size of the book an author does not increase his chances of includ- ing the significant words with which we deal. On the other hand, it would appear that by enlarging the grade offerings he may increase his chances of including such words. The apparent contradiction, however, is easily explained. Our record of the number of words in a book is net--i. e., all duplicates anywhere in the book excluded. On the contrary, our representation of grade offerings (Table VII) shows each word that occurred in a grade regardless of whether or not it occurred in another grade. Thus, it would be possible for a single word to be counted seven times in a book offering instruction to Grades 2 to 8 inclusive. Tc the extent that words thus occurring in more than one grade tended to be Ayres-Teachers College words, enlargement of grade offerings would be positively related to the appearance of the words of this list. Now, it is precisely the words of this list which, more strongly than words outside of it, reappear in successive grades. Hence we find that the size of the grade offerings is posi- tively correlated with the number of Ayres-Teachers College words presented in each grade— and this is despite the fact of absence of correlation for entire - 50 - books when the basis is the net number of words per book. The distribution of the learning and reviewing load as far as the Ayres- Teachers College words are involved is an important consideration for the grade teacher. Table VII shov/s an increase and then a decrease in the number of these words with advancing grades in all the spellers except Books VI and IX. Book VI decreases uniformly from the beginning, while Book IX is irregular. The fact that the number increases and then decreases is due in part to the fact that a given word may appear in more than one grade. For example, it may appear in Grade 2 for the first time and again in successive grades, perhaps Grades 3 and 4. This statement is based on a study of the appearance of 100 words in the different grades. Therefore, the fact that the number of Ayres-Teachers College words increases and then decreases with advancing grades does not necessarily mean that the number of new words from this list increases and then decreases. The data in Table VII, however, seem to indicate that the learning and review of the Ayres-Teachers College words is cumulative up to the end of the middle grades. Hatio of Ayres-Teachers Co liege Words to the Different Words in Each Grade The preceding discussion was concerned with the absolute number of Ayres-Teachers College words, but in a study of the vocabularies of the spellers we are also concerned about the relative proportion of the number of the "most frequently used words" and of the other words not so frequently used. That is, "What percent of the words taught and reviewed in a given grade are the 'most frequently used words'?" Table IX shows the practice for the ten spellers used in our project. Following the data for each grade through each book, one observes that with advancing grades there is a decreasing percent of all the words in the grade that are Ayres-Teachers College words. That is, as the pupil advances through the grades, he is confronted with a less and less proportion of "most frequently used words." This practice can only be justified on the assumption that the - 51 - table IX PER CENT OF WORDS IN EACH GRADE THAT ARE AYRES- TEACHERS COLLEGE WORDS Grade Book r 1 law I II III IV V VI VII VIII IX X 2 82.7 70.3 90.0 93.1 68.6 71.0 Ti H 88.3 3 8o.7 60.5 76.9 90.1 77.6 81.7 64.5 67-3 55.5 65-7 4 70.0 55.3 60.5 78.5 51.0 50.6 63.0 59-1 48.4 57.7 5 80.8 48.3 58.0 71.2 51.8 47.4 60.8 57.1 62.3 55-5 6 63.9 43.8 60.7 64.7 43.8 38.7 56.6 51.7 45.3 49.0 7 55.7 37.1 34.0 50.7 43.0 33.4 54.6 51.7 39.7 47.2 8 29.7 30.7 31.5 39.6 38.4 32.2 33.3 48.8 31.5 48.9 (a) These books do not provide spelling for Grade 2. Table reads: In Grade 2, Book I, 82.7 percent of the different words are common to the Ayres-Teacher s College List; in Grade 3, Book I, 80.7 per. cent of the different words are common to the Ayres-Teacher s College List; etc. - 52 - pupil has learned to spell accurately the larger proportion of these words in the lower grades. Possibly this is true to a certain extent for about 1000 of the words in the Ayres-Teacher s College list, that is, for the Ayres so-called "thousand commonest words." However, when this list is learned there still remain 2000 of the "most frequently used words" to be taught, and it seems scarcely possible that these 2000 words would be anywhere nearly exhausted in the lower grades. In fact, we know from testing children in spelling these words, that they are not taught to such a degree of perfection in the lower grades as to justify dropping them to a large extent in the upper grades. That there is no uniformity of practice in regard to the proportion of all the different words in the grades that are Ayres-Teachers College words is shown by the range of percents in a given grade for the ten bocks. For example, in Grade 2 the range is 14.5 percent, while in Grade 8 the range is 19.2 percent. The greatest range is in Grade 3 where it runs from 55.1 in one book to 90.1 in another or a range of nearly 35 percent. The median of these ranges is 26.0 percent . Percent of the Ayres-Teachers College 3000 Words Found in Each Grade The percent of the three thousand Ayres-Teachers College words in each grade will also give some notion of the teaching load per grade as far as these words are concerned and again of the wide variation in practice among the spell- ers. The data on this basis for our ten spellers are shown in Table X. Again, the range in the percent of the Ayres-Teachers College words is wide, indicating little agreement as to the number of these words that should be taught and reviewed in each of the grades. It should be noted that in Book III Grades 7 and 8 the percents fall low. This is no doubt due to the fact, as mentioned before, that certain long literary selections were not considered. . ■ . •: . , O . • t - A . ' . *< ' ' ' v r - . - 53 - table X PERCENT OF AYRES- TEACHERS COLLEGE WORDS IN EACH GRADE Grade Book I II III IV V VII VIII IX X 2 10.6 16.4 11.1 13.0 (a) (a) 13.2 13.5 (a) 12.3 3 15.8 17.5 19.2 27.9 24.6 27.8 17.7 24.4 19.0 14.5 4 20.3 17.5 26 . 4 32.2 21.0 23.4 20.9 27.2 18.3 14.6 5 26.0 14.9 28.2 29.8 27.4 20.8 22.7 26.9 22.8 13.9 6 21.9 15.5 30.6 27.9 27.3 15.9 21.5 25.6 20.6 12.5 7 14.7 14.1 7.5 14.5 28.1 15.2 21.1 27.2 22.7 11.3 8 12.3 14.4 14.1 14.1 23.9 15.0 9.4 26.0 20.1 l8.6 (a) These books do not provide spelling for Grade 2. Table reads ; In Grade 2, Book I, are found 10.6 percent of the 3000 Ayres-Teachers College words; in Grade 3, Book I, are found 15.8 percent of the Ayres-Teachers College words; etc. - 54 - The percent of the Ayres-Teachers College words decreases in general with advancing grades in Books IV and VI. In Books VIII, IX, and X the percents remain practically the same with advancing grades. In the main for Books I, II, III, V, and VII the percents increase slightly up to the middle grades and then in the upper grades show a slight decrease. This last tendency would seem to be the best because naturally there would be a piling up of these words to provide for automatic learning. The large percent at the beginning of the grades is due to the necessity of teaching the words that the pupils need in their written work and to the fact that these words are found in abundance in the Ayres-Teach- ers College list. A continually decreasing percent assumes that each grade dis- poses of a considerable number of the words; and this can only be done by teach- ing them substantially to the point of perfection. This assumption is too optimistic because we know from the results of testing the spelling of the pupils throughout the grades that they are slow to learn to perfection the spelling of the words that society demands they know. It is possible that the percents in the lower grades are taking care of the proper proportion of the words of this list, but in order that more words of the list might be taught we suggest a larger percent in the upper grades. Therefore, perhaps making the percents in the several grades approximately the same instead of decreasing them with ad- vancing grades would be a better arrangement. Percents of the Ayres-Teachers College Components in Each Grade The extent to which makers of spellers use the Ayres-Teachers College words may be examined from another point of view, namely, the proportion of the words in each of the two components of this list that appear in each grade in each speller. We prepared tables similar to Table X for the percent of the Ayres "thousand commonest words" and for the percent of the Teachers College "second and third thousand most frequently used words" which appear in each grade in each book. We are not presenting these in detail, but we are giving > . ■ • , . . 1 ; * '• * " *- • . > j ■ ■ ; % • - .L . *.* . . . • *. . ' . . , . . . ...... . . . . ■ • . ... * ^ • • ' , . ■ -l ■ ' X'A ■ ' ■ . ' ... ; ■ V - 55 - a composite of the ten spellers for each grade by showing the median percents. The standards of present practice thus obtained are shown in Table XI. TABLE XI MEDIAN PERCENT OF EACH OF THE COMPONENTS OF THE AYRES- TEACHERS COLLEGE LIST IN EACH GRADE Grade Median percent of Ayres List Teachers Col- lege List 2 27.1 6.0 3 37.8 11.9 4 39.6 12.9 5 43.5 15-7 6 38.1 14.9 7 25.6 12.9 8 23.2 12.5 Table reads: In Grade 2 the median on the basis of the ten spellers used in our project is 27.1 percent of the Ayres ’’thousand commonest words” and 6.0 percent of the Teachers Col- lege ’’second and third thousand most frequently used words”; etc. It is obvious that a far greater percent of the Ayres words are in evidence than of the Teachers College words. This is due to two reasons at least. First, the Ayres words having a much higher frequency, have greater chances of being presented in dictation and more obvious value for presentation J in columns. Second, the Ayres list has been known much longer than the Teachers College list as such. - - CHAPTER V GRADATION OF VOCABULARY After the vocabulary has been selected for the pupils of the element- ary schools , the words must be assigned to the different grades. There are many different ways in which this gradation of the selected vocabulary may be done. Starch 1 suggests the following possibilities: "(1) We might distribute the words according to their frequency, putting the most frequent words in the lower grades and the less frequent words in the upper grades. (2) We might put each word into the grade in which children first begin to use it frequently in their writing. (3) W r e might put each word into the grade in which, according to the consensus of competent judges, such as teachers, say it ought to be put." He considered the last method the proper one when he constructed his word lists for the grades. Horn lists the following principles of gradation which "have been most urged of late": (1) "Those words which are most difficult should be pre- sented in the more advanced grades." (2) "Those words which are commonly used by the children in a given grade should be placed in that grade." (3) "Those words which are most commonly used in correspondence should be taught earliest, on the ground that the school must insure the correct spelling of such words before the elimination of pupils becomes serious." (4) "Those words needed in other subjects should be taught in appropriate grades." (5) "Words should be graded according to their length," (6; "Words should be graded according to phonic progression," 1 Starch, Daniel. E du cat i onal p sy ch o 1 o gy , New York: Macmillan Company, 1919, p. 330. 2 Horn, Ernest. "Principles of method in teaching spelling as derived from sci- entific investigation." The Eighteenth Yearbook of the National Society for the rc/pp HublTrS c Ko o T Fu E li sh i r- "Company^ - 57 - TidymarW says: "We are probably agreed that the use of the word is the most important factor determining the grade in which it should be taught. The child should learn to spell words at approximately the time that he is going to use them. The question of the difficulty of the word as a principle of grading / is secondary to that of use. Naturally, a word should not be placed in a grade when it presents little or no difficulty. In general, a word should be placed in the grade where it is used and where it needs to be taught." These quotations contain six principles of gradation which might be used in placing words in the different grades. Those principles are : (1) Place the "more frequently occurring words" in the lower grades; (2) place the words commonly used by the children in a given grade, especially those appearing for the first time in use by the children; (3) place the words in the grade in which ^he opinion of supposedly" competent judges, e.g, teachers, say they should be placed; (4) place in a grade the words that are needed in other subjects in the S- Ver grade; (5) place the words in the grades according to phonic progression; and (6) place the words in the grades according to grads difficulty of spelling the words, the less difficult words in the lower grades and the more difficult words in the upper grades. The principle of length of words is closely correlat- ed with the principle of difficulty as shown by Ayr es^ who found a Pearson coeffi- cient of correlation of C, 73 between the length of the words in his scale and the difficulty of spelling the words. ’.Ve did not attempt in our study to determine the grading of words in the spellers according to the principle of "need in other subjects", or the prin- ciple of phonic progression", because in the first place we did not have reli- able data on either and in the second place the problem offered too many complex- Tidyman, W.F. The teaching of spelling , Yonker-cn-Hudson; World Book Company, 1919. p. 10. Ayres, L.P, Measurement of ability in spelling , Russell Sage Foundation. 1915, P. 38. “ — - 58 •* itiee for our purposes. Neither did we attempt to study the grade placement according to the principle of "opinion of competent judges’* because the list of words so determined is limited in number and lacked quantitative expression for frequency and for difficulty — two factors that v/e wished to use in many differ- ent ways in our project. V/e do not deny the validity of the principles of "fre- quency of use in adult life" and "in child life." They should no doubt function in the gradation of words and we have made certain uses of the first of these two principles in our project. However, it is quite possible to determine the placement of words, with certain limitations, in the grades according to those principles and we feel that this should be done, but we were not able to do so on account of lack of time after we had made our study of gradation by another principle. We confined our study of the gradation of the vocabulary to the prin- ciple of "difficulty of spelling" which is recognised by Horn and by Tidyman as a principle that should operate. It appeals to us as an important principle be- cause in the first place we have three thousand words whose difficulty is ex- pressed in quantitative terms; in the second place, we hold that the words in a given grade should have some relation to the abilities of the children in the given grade and only by using the principle of difficulty could we study this relation, further, the principle of difficulty is operating in grading words because of the wide use of such spelling scales as Ayres's, Ashbaugh’s and no doubt soon the Teachers College sixteen spelling scales. I In our study of the gradation of the vocabulary, we are concerned with (a) the difficulty of the words in each grade in each speller, (b) the grads difficulty as based on the median difficulty of each grade for the ten spellers used in our 3tudy , (c) the difficulty of each speller as a whole, and (d) the . possiblity of ranking each speller on the basis of difficulty. It is impossible at present to determine the difficulty of all the - 59 - words found in the different spellers because of the lack of data on so large a number of words as this would require. We are compelled, then, to make use of the principle of sampling of the word-content of the spellers and to assume that the words taken by the sampling process are representative of all the others in the spellers. The larger the number of words taken by the sampling process and the more nearly the words are taken at random, the more nearly will the data be representative of the true spelling difficulty of the words contained in the books. Our random sampling ’ms confined to taking the words in the spellers which were common to our list and for which we knew the difficulty of each word. As far as the number of words is concerned we were again confined to the three thousand in the combined Ayres and Teachers College lists. We have already shown that approximately two thousand of these words are common to each of the different spexlers. This would seem to be a sufficient number to give a fairly accurate picture of the difficulty of the words in the spellers. The difficulty of the words in these two lists is not expressed in the same units, and it was therefore necessary for our purpose that this be done. Our work was greatly expedited by the conversion table devised by the authors of the Teachers College list'*. They had already expressed the difficulty of the ’./oi ls in uheir list in units of one-tenth sigma from the Buckingham zero of spend- ing difficulty. They then derived by experiment, on the same basis, the value of each column of words on the Ayres Scale for spelling ability. Using these data, we assigned to eacn word of Ayres list its appropriate difficulty figure. Then the Ayres words were typewritten in alphabetical order together with the diffi- culty of each word and its frequency of occurrence in life. The Teachers College words were already printed in alphabetical order showing the frequency and the difficulty of each word. We kept the lists separate in order that we might study sp el 1 i_n& s cal e s , Teachers College Record, 21: 3*5, September, 1920. 5 - 60 - the findings for the different lists. When we wished to study data derived by the use of the two lists, it was a simple and straightforward matter to combine the two set3 of data. Distribution tables were prepared showing the difficulty of the Ayres words and the frequency of occurrence of these words in each grade in each of the ten spellers. Similar tables were prepared for the Teachers College words. Thus we obtained tv/o tables for each grade--one on the basis of the Ayres so- called "thousand commonest words" common to the grade, and one on the basis of the Teachers College so-called "second and third thousand most frequently used words." Then for certain purposes we combined the two tables for each grade in- to one table. These tables were in reality correlation tables for frequency of occur- rence of words in the spellers and the difficulty of these words. On the dis- tribution of difficulty side of each table we found the median difficulty. The data obtained in this manner form the basis of one part of this chapter. Data somewhat similarly derived form the basis of another part of the chapter. In this chapter we shall report the median difficulty of each grade in each book, the median difficulty of each of the grades based on a composite of th“ ^ erl spellers, the difficulty of each of the ten spellers as a whole, and a judgment of the spellers in terms of difficulty. Table XIT expresses the median difficulty of each grade-offering in oacb book as determined by the median difficulty of the words in each grade com- mon to the Ayres-Tsachers College list. A study of the difficulty of each grade in each book reveals that in general, as would be expected, the median difficulty of the words common to the Ayres-Teachers College last and each grade increases »vith advancing grades. That is, if the words common to each grade and the Ayres-Teachers College list are re- presentative of the difficulty of words in the grade, we can say that in general u , .. - ' • • , . . . 4. v . • ' - • « r ■ . h . '1 - 61 - , ^ TABLE XII MEDIAN DIFFICULTY^ OF THE AYRES- TEACHERS COLLEGE WORDS AS FOUND IN EACH GRADE AND BOOK Grade Book I II III IV V I VI VII VIII IX X 2 26.8 32.4 39.2 29.5 (b) (b) 32.2 30.5 (b) 35.7 3 32.1 35 . s 32.7 33.6 34.1 34.4 33.6 33.7 33.1 34.6 4 33.9 36.9 35.1 35.0 35.8 37.5 37.6 35.6 35.3 37.7 5 37.5 38.0 36.3 37.9 38.4 40.7 39.2 38.3 38.0 41.1 6 39.3 39.2 38.2 41.0 38.7 44.0 42.5 41.4 40.4 43.8 7 41.4 36.7 47.6 43.2 40.7 45.2 42.4 43.2 41.8 47.5 8 39.6 39.6 46.8 44.2 CM • H * 47.8 47.2 44.1 42.3 43.4 (a) (b) The median difficulty is expressed in terms of ten times the sigma dis- tance from the Buckingham zero of spelling difficulty. These books do not provide for spelling in Grade 2. Table reads : In Grade 2, Book I, the median difficulty of the textbook words common to the Ayres-Teachers College List is 28.8; in Grade 3, Book I, the median difficulty of the textbook words is 32.1; etc. <■ I - 62 - the difficulty of the words increases with the advancing grades. This would seem to be strong evidence that the principle of difficulty has been operating in the gradation of words in our spellers and that we are measuring a gradation factor common to all the spellers. Hence books may be compared on the basis of diffi- culty without fear of measuring some book by a standard which does not apply. However, in certain books there are some exceptions to the general rule that the median grade difficulty progresses with the advancing grades. In Book I Grade 8 the median difficulty is lower than in Grade 7. In Book II Grade 7 the median difficulty is lower than in Grade 6 and it is interesting to observe that the median difficulty of Grade 7 ia also lower than the median difficulty for Grade 4 of the same book. In Book III Grade 8 the median difficulty is lower than in Grade 7. However, the inversion in this case is not large. In Book VII Grade 7 the median difficulty lies slightly below the median for Grade 6. In Book X there are two inversions. The median difficulty for Grade 3 is lower than that for Grade 2, and the median difficulty for Grade 8 falls below that of Grades 6 and This las “ fact > however, is due to the four pages of dictation material unexpectedly introduced at the end of Grade '3. If our measure of difficulty is what we assumed it to be, then it follows that the presence of inversions in dif- ficulty offers strong evidence of lack of proper grading. A comparison of the median difficulty of a given grade and speller with the median difficulty of the same or other grades in certain other spellers will j show some interesting facts. For example, the difficulty of the words in Grade 2 l! , ' ’’ fi of Book X is greater than in Grade 4 Book I and in Grade 3 in all the spellers ex- j cept Book II. The median difficulty of the words in Grade 3 Book II is greater than in Grade 4 in Books I, III, IV, VIII, and IX, and the same as in Grade 4 Book V. i Another view of the lack of agreement in median grade difficulty may be I obtained by noting in Figure 2 the ranges of grade difficulty as found in the ten 1 spellers. I - . r . i . 'I . • J .. i ' t . . - , . .. i ' f , ..... all e ■ « . ■ . V fc . ( . < " i . • < • . i - 63 - 47 46 45 44 43 44 m 40 39 33 x L 37 'k 36 ^ 35 34 33 32. 3/ 30 29 28 M 'O * V5 h ■ V5 K ^5 | figure £ /Range of Median Grade Difficulties for Jen Spelters da sed on A ynes - 7ea c dens Co/iege fVon/s Grade 4 5 7 F - 64 - Median grade dif ficulty o f the Ayres words . -.Because of the wide uoe of the Ayres "thousand commonest words" as testing words and consequently as teaching words, we decided to express the difficulty of these w'ords that are com- mon to each grade in each book. While we recognize that these words do not give as true a picture of the difficulty of all the words in a given grade as does the Ayres-Teachera College list of which it is a component, they will neverthe- less give an added touch to the picture. Furthermore, since, as v/as shown in Chapter III, all of the spellers contain a very large percent of these words and because each grade contains a relatively high percent of them, we can observe fairly accurately the way in which the authors have graded the Ayres words. Tabic XIII, therefore, shows the median difficulty of the words in each grade that are common to the Ayres list. It will be seen that in general the median grade difficulty becomes greater as the grades advance. Nine exceptions occur out of fifty-seven possible cases. All these appear in Grades 7 and 8. We are not surprised to find inver- sions when we are dealing with the Ayres words because many of the high frequency words of this list are found in the texts that have dictation work for the pupils since they are words that are necessary in sentence forming. The inversions in the upper grades are no doubt due in part to the fact that many easy dictation words that are sentence builders are needed in order to introduce to the children the mere difficult words which is the real purpose of the dictation material. The ranges are shown in Figure 3 which was constructed on the basis of the data in Table XIII. Difficulty based, on words common to the Teache rs College list. The Teachers College words on the whole have a much lower frequency of occurrence. Hence it is safe to assume that not so many of them are found in the texts sole- ly for the purpose of forming sentences as was the case with many of the Ayres w ? ords. Therefore, the medians calculated on the basis of the Teachers College < ( < ' - •• . r . ... . . ; . . ... 1 . . « i < ♦ * .... - 65 - TABLE XXII MEDIAN DIFFICULTY OF THE AYRES WORDS AS FOUND IN EACH GRADE AND BOOK Grade Book I 1 11 III IV V VI VII VIII \ T J X 2 27.9 30.6 29.2 28.3 (a) (a) 29.3 * u jar4» 29.2 (a) 30.0 3 30.7 32.1 31.0 31.4 32.4 32.4 31.9 31.4 31.0 32.8 4 32.4 32.9 32.5 32.7 32.5 34.0 33.9 33.4 32.9 36.5 5 34.6 33.2 33.4 34.4 34.3 38 . 4 37.4 34.7 35.2 40.6 6 35.3 33.1 34.0 36.2 33.8 42.9 41.7 38.8 35.8 43.8 7 36.0 32.8 45.9 38.1 33.9 44.2 38.8 40.4 36.5 47.5 8 33.4 33.4 43.6 35.4 34.4 45.8 43.5 36.3 40.9 34.7 ( a) ' These books do not provide for spelling in Grade 2. Table reads : In Grade 2, Book I, the median difficulty of the textbook words common to the Ayres thousand words is 27*9; in Grade 3, Book I, the median difficulty of the textbook words common to the Ayres thou- sand words is 30.7; etc. • ^ J, *J . words that are common to a grade will, as no doubt many readers will believe, give a better measure of the difficulty of the words intended by the authors to be taught in the different grades. The median difficulty of the Teachers College | words that are common to each grade in each book is given in Table XIV. It is observed again that in general the median grade difficulty in each book becomes greater as the grades advance. The only exceptions are found in Book I Grade 4, Bock II Grade 7 , and Book IX Grade 6, It should be noted that the number of inversions when the Teachers College list is used is less than when the medians were calculated on the basis of the Ayres words. This may indicate that the Teachers College words are easier to grade from the standpoint of diffi- culty. If the assumption is true that the Teachers College words are not needed so much for the purposes of sentence construction as are many of the Ayres words, then we can see how it would be easier to place these words in the proper grade, A few comparisons of median grade difficulty will reveal present prac- tices in placement of words. The median difficulty of the Teachers College words common to Grade 2 Bock II is 37*7 which is greater than the median difficulty in urade 3 in Books I, ICT, IV, VII, and X. The median difficulty of similarly se- lected words in Grade 3 Book II is 40.1 which is greater than the median diffi- culty in Grade 4 Books I, IV, VIII, IX, and X. The median difficulty in Grade 4 Books II and VII is equal to or greater than the median difficulty in Grade 5 Books I, TIT, IV, VIII, IX, and X. In Grade 7 Book X the median difficulty is 47.6 which is greater than in Grade 8 Books I, II, V, and IX, and the same as the median difficulty in Grade 8 Book VIII. One is inclined to feel that the preceding data are illustrations of displacem ent, of words rather than place ment of words. If, for example, one speller has a median grade difficulty of words for a given grade and another uses; the same degree of difficulty of words for a grade or two higher, which one is correct or, more practically, which speller should be selected for the given ir rade? - 67 - 46 45 1 44 43 42 4/ 40 30 23 ^ 37 I T * ,5 <0 * £ \ ,5 i V5 35 34 33 32 . 3 / * V5 V* 29 26 27 M § V5 A/<7 L/re 3 f?ar>ye or Med/a/? Grade B/ff/cu/f/es for 7er? Spe//ers Based 0/7 Byres Words Grade 5 e - 68 - TABLE XIV MEDIAN DIFFICULTY OF THE TEACHERS COLLEGE WORDS AS POUND IN EACH GRADE AND BOOK Grade Book I II III IV V VI VII VIII IX X 2 34.4 37.7 36.2 36.1 (a) (a) 35.9 35.4 (a) 34.4 3 37-5 40.1 37.2 37.5 40.6 39.3 37.6 37.7 38.3 37.1 4 37.1 41.5 40.4 39.2 41.0 40.9 41.5 39.4 39.0 38.7 5 41.3 42.7 41.8 41.4 42.5 42.3 43.1 41.6 41.2 41.4 6 43.7 44.2 42.6 43.3 43.0 45-0 44.0 43.1 39.4 43.8 7 44.9 43.0 48.9 45.4 44.8 46.3 45.4 45.1 44.0 47.6 8 45.6 45.4 49.2 48.6 45.1 49.4 48.9 47.6 44.8 49.2 (a) These books do not provide for spelling in Grade 2. Table reads: In Grade 2, Book I, the median difficulty of the textbook words common to the Teachers College List is 34.4; in Grade 3, Book I, the median difficulty of the textbook words common to the Teachers College List is 37*5; etc. J Another view of the extent to which the material in a given grade and book is similar in difficulty to the material in the same grade of a difforent book may be seen in the series of grade ranges as shown in Figure 4. One of the most striking facts is the grade overlapping. For example, there are some spellers whose median difficulty for Grade 3 is as easy as or eas- 1 ier than that cf Grade 2 in other spellers. The extent of "overlap" for Grade 3 on Grade 4 is great. The range for Grade 5 Is relatively narrow and "overlaps" slightly cn Grade 4. Grade 6 completely "overlaps" Grade 5 and extends over the uppei half of the range for Grade 4, it is the grade in which the textbook mak- ers are most at variance with each other. Grade 8 "overlaps" to a great extent Grade 7 and the upper limit goes relatively only a short distance above the upper limit of Grade 7 . Another way of looking at the difficulty of the grades expressed in terms of the Teachers College words common to them is to express the difficulty coefficient in an equivalent column and word found in the Ayres scale for meas- uring spelling ability. Table XIV-A shows these data. In the first place we have shown for the median grade difficulty that ^.here is progress in difficulty with advancing grades, whether we use the compos- ite Ay res -Teachers College list, the Ayres list of the "thousand commonest words’,’ cr the Teachers College list of the "second and third thousand most frequently used words." We have shown that in certain spellers there are exceptions to the general rule of greater difficulty with advancing grades regardless of whether we I . used as a basis for deriving the median difficulty the Ayres-Teachers College list, the Ayres list or the Teachers College list. We have shown also that a higher grade in a given speller is not necessarily more difficult than a lower grade in another speller. For example, the median difficulty in Grade 5 Book I \ j is less than the median difficulty in Grade 4 Book VII on the basis of the Teach- ers College list. Further, we have shown the extent of the distribution for the ! - 70 - 49 48 47 46 45 K * V5 i V5 44 43 42 || 40 ^ $ 3d § vs Ijh ,5 vs 5 Vs 37 36 3S L24 03 3Z b/ 30 M £ vs I M/p ore 4 Marpe or A/ec/( a /? Grade Mjffrco/t/es for 7 e/? Spaders Zased or Teachers Co//ege M/ords Grade 0 6 - 71 - TABLE XIV-A THE MEDIAN DIFFICULTY OF THE TEACHERS COLLEGE WORDS AS FOUND IN EACH GRADE AND SPELLER EXPRESSED IN TERMS OF COLUMNS AND WORDS ON THE AYRES SCALE Book I II III IV V VI VII VIII IX X 2 M N N N (a) (&) M M (a) M trust except except except trust trust trust 3 N P N N P 0 N N 0 N except spend except except spend eight except except eight except 4 N P P 0 P P P 0 0 0 except spend spend eight spend spend spend eight eight eight 5 P Q P P Q Q Q P P P spend final spend spend final final final spend spend spend 6 Q R Q Q Q R R Q 0 Q final lose final final final lose lose final eight final 7 R Q T R R S R R Q T lose final guess lose lose often lose lose final guess 8 S R U T R 'U T T R U often lose meant guess lose meant guess guess lose meant (al These books do not provide spelling for Grade 2. Table reads : In Grade 2, Book I, the median difficulty of the Teachers College words common to it is equivalent to Column M on the Ayres spelling scale and to the word trust found in this column; etc* - 72 - ten spellers cf the median difficulty for each grade in each speller. The inversions within a speller and the varying difficulty of the words in a given grade in the different spellers make the problem of selecting a spell- er to suit the needs of the children in the grades of a given school very diffi- cult. First we should know as accurately as possible the spelling abilities of the pupils in the different grades for whom a text in spelling is to be selected and second we should know the difficulty of the words in each grade of the spell- ers under consideration. With these tv.'o factors known the problem resolves it-, self into deciding which speller is best suited to the needs of the pupils of the different grades. As we see the problem in the light cf the varying difficulty within a speller from grade to grade and between spellers for a given grade it is scarcely possible that one text in spelling would be suited best to the needs of the children in several grades. It is possible to find a speller in which the words are suited to a given grade; but when we consider the needs of the children in the seven grades (2 to 8), and attempt to find the one speller in which the words best suit the needs of the children for all grades, we have un- dertaken a difficult problem. Therefore, it would seem that under present cir- cumstances a school system would have to select more than one speller. This leads to the further conclusion that spellers should be published for the sepa- rate grades rather than one speller for all the grades. Difficulty by Grades , What is the difficulty of the soeiling material that is being given to ! ! the children in each grade in our American schools today! In order to find an answer to this question an index number was computed for each grade by calculat- ing the median difficulty for a given grade in all the ten books. We are assum- ing that the ten books used in this study are typical spelling texts in our American schools today and that the words that are common to our standard list and the textbooks are typical of the words that are in these standard textbooks. J . .it.:*; 1 •/ • . , ■ *’ ■ • ... . •- . ... ~ L ' '• : * * .i. ‘ , - ' ■ ' t . .. .. t • * , J" .. >/. • - •- • . . i . •. ■ - -,i ■ i t ■ k '- 1 ' 1 *. • 1 , I . - 73 - A difficulty distribution table was constructed for each grade and the frequency of occurrence of each unit difficulty in a given grade for all the hooka | were added in order that a total frequency for each unit could be found. This prc-J cedure gave eeven tables— one for each of the grades from 2 to 8 inclusive. A series of seven tables constructed in this manner was formed for the words that were common to each grade in each book and to the Ayres-Teachers Colley i list. Another series was formed for the words that were common to each grade in each book and to the Teachers College list alone. Thus we calculated two sets of inoex numbers for each grade from grades 2 to 8 inclusive. The index numbers calculated on the basis of the words that were common to the Ayres-Teachers College list and tc the respective grades in the ten books are shown in Table XV. TABLE XV INDEX NUMBERS BY GRADES A.-T.C. 1S0RDS Grade Index Number Grade Interval 2 30.4 3 33.7 3.3 2.1 4 35.8 5 36.6 2.8 6 41.2 2.6 7 42.9 1.7 8 43.7 0.8 1 - I- 1 Tbs table read e;; The index of difficulty for Grade 2 is"~3C.4; the index number for Grade 3 is 33.7; and the interval between Grades 2 and 3 is 3.3 unit distances; etc. The difficulty index number increases, of course, with advancing grades, but the steps between the grades decrease in size from the lower grades to the higher ones. The range from the index number for Grade 2 to the index number for : | Grade 8 is 13.3 units. Since these figures indicate difficulty measured from zero, it is ap- i parent that typical eighth-grade words are less than half as hard again as typ- ical second-grade words. The range of 13.3 seems small. It is less for example - than the range for a single grade according to the practice of different books (See Grade 7» Figure 3)* We shall have occasion in the next few pages to examine the range be- tween grade medians from different points of view. We shall find that from all these points of view the ranges are relatively small. We have in other connec- tions pointed out to the point of tediousness the large degree of variation which is exhibited among things which purport to be the 3 ame. Here wo have on the contrary a relatively narrow variation among things which purport to be dif- ferent. As we shall point out later, this leads to a situation in which the higher grade offerings of the text books fall far short of the ability of the pupils in the same grades. Again in order to get a picture of the difficulty of the grade material uninfluenced by the many sentence-forming words of the Ayres list, it was de- j cided to calculate the grade index numbers on the basis of the words that are common to the Teachers College list and to each of the separate grades in all the books. Table XVI gives the results of these calculations. - 75 - TABLE XVI INDEX NUMBERS BY GRADES T.C. WORDS The table reads; The index cf difficulty for Grade 2 ia 35*9; the index of diffi- culty for Gradd 3 is 38.3; ana the inter- val between Grades 2 and 3 is 2.4; etc. The range this time is from 35 .? in Grade 2 to 47.3 in Grade 8 . This gives a distance of 11.4 units for the seven grades. The two largest intervals are between Grades 2 and 3 and between 7 and 8. The intervals between Grades 4 and 5. 6 and 7* end 7 and 8 are practically the same. j Comparing the index numbers computed on the basis of the Ayres-Teach- j ers College List (Table XV) with those computed on the basis of the Teachers College List alone (Table XVI), one finds that the latter are larger grade for j grade without exception. Further, it will be noted in comparing the grade in- | I tervals shown in the two tables that whereas there was in the one case a con- eietent decrease in the intervals for advancing grades, there was in the other | case irregularity in the behavior cf the intervale. This statement is shown j clearly in Figure 5 which shows the curve for the grade intervals based on the j A.-r.c. words and the one based on the T.C. words. From the best evidence that l we have about the spelling ability of the children in the elementary schools. j •we feel justified in saying that the intervals that were computed on the basis 5 Grade Index Grade Number Interval 2 35.9 2.4 3 33.3 1.6 4 39.9 M 5 41.8 i 1.6 6 7 43.4 1.9 4 5* 3 2.0 8 47.3 ! * Grades \ [ j~ ; | *p\| S-6 \ ~ 6 \7 ! t{& figure 5 Grade Jr ferya/s /?efwee/7 Judex A/uru6er& X-rc. - Xyses- 7&<2cX>e/'S e//// a //■ £z 3 X -1 39.9 The easiest book according to these index numbers is Book I and the most difficult one is Book VI, The easiest book is as easy as the words of Col- umn M on the Ayres Scale for Measuring Spelling Ability and the most difficult book is as difficult as Column P on the same scale. In terms of grade ability Book I has a median difficulty equivalent to approximately half way between Grade 3 and Grade 4, while Book VI has a difficulty slightly above the median ability of Grade 4. The reader should not forget in making the comparisons that we are dealing with medians in both the difficulty of words and in the spelling ability of the children. j - 82 - The index numbers that have been given were calculated in such a way that the difficulty of the words in the different grades were taken into consider- ation. This method of calculation gave extra weight to the words that were re- peated more frequently from grade to grade. In order to place the words on a more nearly equal footing in this regard, .it was decided to calculate the median diffi- culty of each book from the distribution of the Ayres -Teachers College words after all the words in a book had been alphabetized and duplicates had been eliminated. This method does away with the extra weight that was given to the words that crossed grade lines in the preceding method of calculation. The median difficulty obtained on the entire distribution of the Ay res -Teachers College words is given in Table XIX, TABLE XIX MEDIAN DIFFICULTY OF EACH BOOK, SECOND METHOD. pr— Book Index Number Bo©k f"“ — Index Numb er I 39.7 VI 41.3 II 40. 2 VII 40.4 j j III 40.7 VIII 40.4 IV 4©, 6 XX 40.3 V 40.9 — — — — — — X 1 40.5 According to these index numbers Book I is again the easiest and Book VT is again the most difficult. But it will be observed that the second method gives higher index numbers than the former. This is no doubt due to the fact that the relatively easy words cross grade lines more frequently than the more difficult words. These index numbers give Book I about the same difficulty as the words of Column 0 on the Ayres Measuring Scale for Ability in Spelling, while Book VI has difficulty a little higher than the words of Column P on the same scale. A comparison with the median grade ability as computed by Buckingham shows that both these books and consequently all of them are slightly below the median ability for Grade 4. The index of difficulty was computed on smother slightly different ba- sis. This was done in order that a large part of the high frequency words might be omitted. This also takes out of consideration many of the relatively easy words that are used because of the necessity of making sentences in those books that had dictation work for the pupils. Many of these words were given by the authors not because they were words offering spelling difficulty but because they had to be used to make sentences. For each book the word showing was the number of different words, without duplicates, common to the book and to the Teachers College list. Table XX gives the index numbers for the entire distribution found by using these words. TABLE XX MEDIAN DIFFICULTY OF EACH BOOK, THIRD METHOD Beck Index Numb er Book Index Number | 1 42.3 VI 43.6 f n 42.3 VII 42.9 | III 43.3 VIII 42.6 IV 42.9 IX 42.4 V 43.2 j X 42.5 Again Book I is the easiest and Book VI is the most difficult. It will also be seen that these index numbers are higher than the two preceding. In comparison with the Ayres Scale, Book I stands about half way between Columns j P and Q, while Book VI stands slightly above Column Q. In comparison with the j Buckingham median grade ability, all of the books are above the median ability of Grade 4, and Book VI, the most difficult one, stands about half way between the median grade ability of Grades 4 and 5* So far as index numbers for the books are concerned, we are not able - 84 - to say which book has the optimum one. We do not know what should be the index number in an absolute sense. But by reasoning we can arrive at an approximation. If a given book ia to be used by children from Grade 2 to 8 inclusive, we can take the middle grade and say that the book should have a difficulty index that approximates the ability index for that grade. Applying this standard, we find that none of the books is up to the standard. All are too easy, because none i eacnos 46.6 which is the median ability for Grade 5* Here we are compelled to recognize that the social standard for spelling is perfection rather than median p erf orman ce. C^n£ ura_s^n_ 0 j : _y 2 ^_^^ 0 jis. _on the basis of dif ficulty « — -It is possible to compare the spelling books on the basis of difficulty. In order to do so certain assumptions will be made. First, we shall assume the median grade abilities as defined by Buckingham? are substantially correct. Second, we shall assume that the medians computed by us from the distributions of the Teachers College words in the grades are the most nearly true medians for all the different words that are in the different books. This assumption is based (a) on the fact that the medians computea on the Ayres-Teachers College words are lower than the grade medians should be because of the large number of easy high frequency words, and (b) on the fact that many words in the books are more difficult than those of our last, Honce the chances are that the medians computed on these words — i.e. the Teachers College words that are somewhere between the relatively easy words fo'i which we have data and the relatively difficult words for which we do not have data— give a more nearly true picture of the difficulty of the texts. Third, we shall assume that the gap between the spelling ability of the pupils in the grades and the difficulty of the words in the textbooks for the grades is too wide. Hence the grade difficulty of all these books should approa- ch more nearly the median grade ability and among these spellers the best will 7 ~ ’ ~ — — - — — Buckingham, Ibid. p. 61. be the one whose grade difficulties differ in general from these median grade abilities by the smallest amount. We are safe in making this assumption because none of the books approaches very closely to this standard except in Grade 4 whore the two medians are equal for all practical purposes. There should always remain at least in the higher grades a gap between the median grade ability and the median difficulty of the words designed for a given grade--!. e. pupil ability should be higher than word difficulty. This is due to the fact that society demands that the pupil performance should be per- fection rather than median performance. It should not be forgotten that the a- bixity factor is the same as the performance. On the other hand in the lower grades (perhaps Grades 2 and 3) the pupil ability should fall below word diffi- culty because the children in these grades have been taught comparatively no worth before this time. If at this point the reader will refer to Table XIV he will have before him the best determination we can offer of the difficulty of the material pre- sented in each grade for each book. A summary of the details of Table XIV yields for each grade (all books combined) a difficulty figure. This figure represents as nearly as our data permit, the current practice of textbook writers with re- spect to grade difficulty. This summary may now be brought into relation with the ability of the pupils in these grades. Table XX-A dees this. The first col- umn of this table is taken directly from the second column of Table XVI. The entries in the second column are the figures of grade ability derived from p Bu ck ingham . Figure 7 shows the same data as does Table XX-A. The full line re- presents the average difficulty of the spellers by grades (column 2 of Table XX- A) the dashed line represents pupil ability by grades (Column 3 Table XX-A). Q " ' ' 111 ’ " “ " Buckingham, Tbid p. 61, «t j, V I . •• l . . J 1 * n • '<■ ' . . .> •> ; . : ,u : •: - : j.J . • v . . : . . \ . ■ : . . . r ■ V ' - 86 - In addition to these curves we have inserted a third ere which is entirely theo- retical. It is the curve which lies between the difficulty and ability curves. This curve represents our judgment of a better adjustment of grade offerings to pupil ability than is exhibited by any of the spellers as a whole that we have examined. We are not disposed to defend the exact position of our theoretical curve as shown in Figure 7* The precise placement of this line must be left tc future experiment. We submit, however, that present spellers are too hard in the lower grades and too easy in the upper ones. Some sort of "in-between” posi- tion must therefore be taken by a curve which would represent better conditions. However, in order that our theoretical curve might not be wholly sub- jective cn our part, we located it on the basis of actual practice as found in our ten spellers. For example, the position of the curve for Grade 2 is the grade difficulty taken from that speller in which the difficulty of Grade 2 has the least discrepancy from the corresponding grade ability as determined by Buckingham, It will be noted that according to Figure 7 the ability of Grade 2 lies below the median difficulty of the Teachers College words. This is to be expected because the children of this grade have learned very few words up to this time in their school career. A test of their spelling ability would show that they were able to spell only vein- few words and that these words were very easy. The whole field of spelling is ahead of them. When they enter the next grade they will have learned many words that they have been taught, and hence they will show a much higher ability in spelling. Figure 7 pictures this quite clearly. When the pupils enter Grade 4 they will have a median spelling ability that approximates the median difficulty of the words that they are taught. In the succeeding grades the median ability draws farther and farther away from the median difficulty of the words that they are being taught. - 87 - TABLE XX -A COMPARISON OF MEDIAN DIFFICULTY WITH MEDIAN PUPIL ABILITY BY GRADES. Grade Difficulty (all books combined) Ability of Pupils (af- ter Bucking- ham) Difficulty - ("Best” prac- ■ tice) a 2 35.9 16.4 34.4 3 38.3 31.9 37.1 4 39.9 40.5 40.9 1 5 41.8 46.6 43.1 6 , 43.4 53.7 45.0 7 1 45.3 58.1 48, 9 \ 8 47,3 64.3 j, 49.4 a That grade difficulty which is found in one of the ten spellers and shows the least discrepancy from the corresponding grade ability (after Buckingham). Table reads: In Grade 2 the median difficulty of all bocks combined is 35.9. the median pupil ability (after Buckingham) is 16,4, and the "best” practice shows a grade difficulty of 34.4; etc. : ■ , - 88 - 65 60 55 s1/>////y 50 45 40 p5-§ S?^ zs 2.0 /5 /O 2?//y/4 / s/ty,7~A///// &//&/- /SvcA/Oy/ra/?? Grade 4 6 7 8 - 89 - The divergence of the median ability from the median difficulty in the advancing grades is what society demands as a result of the teaching of spelling* In fact society demands that the outcome of the teaching of spelling should be perfection rather than these median performances. This rigid demand goes beyond the capacity of the pupils end causes so much of the pains on the part of the teacher and the taught. The teachers of spelling have tc resort to all kinds of devices to induce the pupils to stick to the task of mastering to perfection the words that are presented to them. From the standpoint of what we know of the distribution of the ability or capacity of the pupils in the elementary schools, v/6 feel safe in assuming that the words are too easy as they are found in the American textbooks on spelling. The pupils feel that they have mastered to a de- gree of perfection that fits their needs of life the words that are presented to them. If the words presented more real difficulty tc the pupils they would study with much less discomfort the spelling that was assigned to them. We do net wish to defend the exact position of the theoretical diffi- culty line in Figure 7. y/e may have misplaced it many units for each grade. We do not know. The placement of this line must be left to future experimentation. With the idea in nind that for each grade the discrepancy between book difficulty and pupil ability (either way ) should be considerably less, v/e found for Book I the difference between the difficulty of its second grade offerings and the ability of second grade children to spell, then we found a similar differ- ence for the »hird grade, and so on to the eighth grade. These differences were then added, giving the aggregate amount of discrepancy between the difficulty grade by grade o f Book I and the abilities of pupils in the same grades. Similar aggregates were obtained for the remaining nine books. However, in practice we excluced the showing for Grade 2 because some spellers did not offer material for this grade. The aggregate discrepancy and the rank for the ten spellers are shown in Table XXT along with the aggregate discrepancy for the speller represents ed by our theoretical curve in Figure 7. The book whose total of differences ■ » <. n - . . . . t , < . . •; ■: * - 90 was the least was given first rank or a rank of 1; the book having least total was ranked 2; etc. The ranking of the ten spellers is Table XXI. TABLE XXI RANK OF SPELLERS BASED ON ADJUSTMENT TO THE ABILITY OF PUPILS Book Discrepancy from Buckingham grade ability Rank III 46.1 1 VI 47.2 2 VII 47-7 3 X 48.2 4 IV 51.4 5 VIII 52.7 6 V 56.2 7 II 56.3 8 I 56.7 9 IX 61.7 10 Theoretical book 42.4 next to th e shown in - 91 - CHAPTER VI REPETITION OF WORDS Psychologists who have investigated the laws of learning involved in acquiring control of the correct spelling of words are generally agreed that the dominant type is habit-formation. The following quotation from Starch 1 is typical of their conclusions: "Skill in s P? Ui ”g is primarily a matter of forming associative con- nections between certain arbitrary symbols arranged for the most part in arbitra- ry order. Economy in the learning of spelling reduces itself to this question: Under what conditions can these associations be made most quickly, most effec- tively and most permanently? Of the four laws of association, frequency, vivid- ness, primacy, and recency, the first two are most directly applicable. Obviously, frequent repetition is necessary to establish the connections. Frequent reviews', monthly, weekly, and possibly daily are indispensable." Students of the pedagogy of spelling have reached the same general conclusion that learning to spell is chiefly habit-formation. The following statement by Tidyman 2 is typical: of " T S ? e i^ nS ’ ® S have seen, involves the association of mental images . ords. Learning of this kind, which has in it less of thought, variety of content, and originality, and more of the repetition of certain set forms^r 15 lean r ing of v the habit-formation type. It consists primarily of the a sociation of certain arbitrary symbols, and then of practice in making these Sf > 2iih UB i il * th ? y ^ be . made aut ° ffi atically, or -without thinking- aboSt them. The emphasis m learning is upon exercise and effect. The principles for the formation of habit are vivid picturing, attentive repetition, and automUc e.n- trol. These principles represent also the steps or stages of learning." Both of the quotations call attention to the two primary laws of habit- formation which are exercise and effect. The law of effect is concerned with the vividness with which the picture of the word is presented to the learners. The word may be written on the blackboard, underlined in the speller, printed in Starch, Daniel. Educational Psychology , p. 343. 2 Tidyman, W. F. The teaching of spelling , p. 36. j 92 - bold face type when first presented, analyzed into stem and prefix or suffix. In short, a countless number of devices are found in modern spellers and are used by the present day classroom teacher. We investigated certain phases of presen- tation in the spellers designed to assist the teacher in making the words vivid, and our report will be found in a later chapter. The law of exercise functions in many ways in spelling. The child in studying a word repeats the letters in order, time and time again, orally and in writing. The classroom teacher requires the children to spell the words oral- ly and in writing both in columns and dictation. All this is done to insure the amount of exercise or repetition that is required to make the correct spelling automatic. Some spellers provide for repetition of words by daily, weekly, monthly, half-yearly and yearly reviews. For example, some spellers use every fifth lesson as a review by printing again the words in the four preceding les- sons; other spellers at regular lesson intervals direct the pupils to review certain preceding lessons; still others print the review words under such captions as, "Spelling Match" or "Review Words"; or again, they may provide for review by reprinting tne words through a series of lessons for one or more grades. All of these devices aim, among other things, to insure repetition until correct spell- ing has become automatic. Owing to the recognized importance of the repetition of words by psychol- ogists, by writers on the pedagogy of spelling, by the practice of better teach- ers, and by the authors of spellers we decided to investigate certain of the provisions for this phase of spelling as exhibited in the ten spellers used in our study. Among the more important phases of repetition investigated were (1) the average number of occurrences per word in the entire vocabulary of each spell- er; (2) the average number of occurrences per word in each grade and in each speller; (3) the repetition of the Ayres- Teachers College words common to each grade; (4) the extent of repetition across grade lines; (5) the correlation -93- between repetition and difficulty of words; and (6) the correlation between repetition of words in spellers and in life. The average number of repetitions per word is shown in Table XXII. TABLE XXII NUMBER OF WORDS, NUMBER OF OCCURRENCES, AND AVERAGE OCCURRENCES PER 'WORD Book Number of dif- ferent words Total number of occur- rences Average number of occur- rences per word Book Number of dif- ferent words Total number of occur- rences Average number of occur- rences per word I 3968 15509 3.9 VI 5328 16624 3.1 II 5001 11594 2.2 VII 4545 10211 2.3 III 4787 19523 4.1 VIII 4767 16034 3.4 IV 4742 12775 2.7 IX 5915 11121 1.9 V 5911 16942 2.9 X 4159 5896 1.4 gable r e ads : In Book I there are 3968 different words occurring 15509 times or an average number of occurrences per word of 3.9 times, etc. The table shows that some spellers provide for a relatively high number of repe- titions per word while others provide for very little. The range in average number of repetitions is from 4.1 in Book III to 1.4 in Book X. The variation in the construction of spellers from the standpoint of repetition is shown sharp- ly by comparing the average repetition per word for Book X which has 1.4 with that of Book VI which has 3.1. These books are practically free from dictation material and the comparisons are direct. It was suggested that dictation material influences the average num- ber of repetitions per word and one would be inclined to say, "The more dictation, the more repetition"; however, Book II, having next to the largest percent of dictation, has an average number of repetitions per word of 2.2 while Book IX, having tne least percent of dictation, has an average number of repetitions per word of 1.9. However, Book III which has the greatest average number of repeti- tions per word has also the greatest proportionate amount of dictation. Date of publication has no influence on the average repetitions per - 94 - word as may be seen by comparing Books I and X. These spellers are the most recently published of the ten in our list, we find that Book I has 3.9 average repetitions, which is next to the highest* while Book X has 1.4 average repeti- tions per word, which is the lowest of our ten spellers. It is obvious that Book X is in a class by itself when we compare it with Books I and VI. The size of the vocabularies in Books I and X is about the same. However, when we compare the total number of occurrences (i. e., the number of printed words) we observe that Book I has approximately three times as many printed words as Book X. This means, of course, that the words in Book I are repeated nearly three times as often as the words in Book X. The explana- tion for part of this difference is due to the absence of dictation material in Book X. A comparison of Books VI and X will aid further in showing that Book X is unique. In the first place, both of the spellers are practically free from dictation material; and in the second place, regardless of the fact that Book VI has 1100 more words in its vocabulary, it has nearly three times as many printed words as Book X. That is, the difference of 1100 words in the vocabularies of the two spellers accounts only in part for the wide difference in the number of printed words in the two spellers. Therefore, Book X is unique in that it has a small vocabulary, a small number of printed words, and very few words are re- peated. Average Number of Repetitions of All Words by Grades In the preceding section we showed the average number of repetitions per word for all words using the whole book as a unit. This gave a general pic- ture of the spellers. In order that we might get a detailed impression of the repetition per word in each grade, we divided the number of printed words in each grade by the number of different words in the grade. This gave the average num- ber of repetitions per word, when this was done for all the grades in all the books, we derived Table XXIII. -I 96 Grade Range 2 4.4 3 2.5 4 2.1 5 2.0 6 1.8 7 1.1 8 1.0 Table reads : In Grade 2 the range be- tween the highest and the lowest average number of repetitions per word in the ten spellers is 4.4 repetitions; etc. yhile these ranges in the average repetitions per word appear small, yet they are significant. For example, if a speller has a vocabulary of 4000 words, none of which is repeated * it will have only 4000 printed words. If, however, it has an average of one repetition per word, it will have 8000 printed In the first place, it may be noted that in general there is a tendency for the average number of repetitions per word to decrease with advancing grades. Especially is this observable in the spellers that have relatively high repeti- tion in the earlier grades. For example, Book I has an average of 4.1 repetitions per word in Grade 2 which becomes less with advancing grades until in Grade 8 the average is 1.4. In the second place, it will be seen that in the upper grades, e. g. Grade 8, the spellers uniformly provide for very little repetition of the words by printing them. The high frequency in the lower grades indicates that the authors made provision for habit-formation in these grades. The range in average repetitions per word for all words is shown in Table XXIII-A which was derived directly from Table XXIII. TABLE XXIII-A RANGE IN AVERAGE REPETITIONS PER WORD FOR ALL WORDS words; or if it has an average of two repetitions per word, it will have 12000 printed words. Regardless of the fact that these ranges in average repetitions seem small, they are, however, significant when we consider them in relation to the size of the entire vocabulary of a speller. There is a level of uniformity in the average repetitions in Books V, VI, IX, and X. The latter speller , as will be noted by following through the grades, provides for practically no repetitions of words. The words are printed once and seldom or never again in the speller, indeed, so seldom is a word re- peated that we feel confident that the repetitions are inadvertent. Only in Grade 3 where dictation material is provided do the repetitions rise to a point comparable with those of other spellers. There is a certain type of provision for review used in some of the spellers that should be considered in this connection. Instead of reprinting tne words, the autnor directs the pupils to study for a given review lesson certain specified previous lessons, e. g., the review lesson is headed "Review Lessons 15—20." In our study of repetitions of words we made no provision for such devices. We admit that this is a type of review, but would it not have a better influence on the mind of the learner if he were not told to "turn back" aad go over something he has studied before? Students of child life have ob- served that pupils are averse to the dictum "Turn back!" in their school work, f’/e fancy that the better procedure would be to reprint the words as they need review. One measure of the relative worth of a speller is the number of repe- titions per word. Within reasonable limits, and other things being equal, the speller having the highest number of repetitions per word through the several grades should be considered the best speller. On this premise, the spellers rank as follows: first, Book I; second, Book VI; third, Book III; fourth, Book V; fifth, Book IV; sixth, Book VIII; seventh, Book II; eighth, Book VII; ninth, . ' 7 ... • • - . •l . : . . .. • ■ v ... . •• i • . • . . ! * . . . • ‘ • ' ’ . ; . i .. .■ . >1 J 1 i ■ .... ; - . ' t ; t -i . : . j, . ' ■■ ■ I • ... . . . • : ; ■ TABLE XXIV AVERAGE NUMBER OF REPETITIONS PER WORD FOR THE AYRES-TEACHERS COLLEGE WORDS IN EACH GRADE Grade Book ammm '■ I II III IV V VI VII VIII IX X 2 4.6 2.1 5-7 2.5 (a) (a) 2.5 4.1 (a) 1.2 3 4.1 2.3 3.8 2.0 1.8 2.1 2.1 3.1 1.9 1.1 4 3-7 2.1 3.5 1.7 2.3 2.4 1.8 1.9 1.6 1.1 5 3.3 1.8 3.5 1.5 2.3 2.3 1.8 1.7 1.5 1.0 6 3.1 1.7 3.1 1.4 2.2 2.2 1.4 1.6 1.4 1.0 7 1.9 1.8 1.7 1.4 2.0 2.3 1.4 1 . 6 1.4 1.0 8 1.6 1.7 1.1 1.4 2.2 2.2 1.3 1.6 1.3 1.7 (a) These books do not provide for spelling in Grade 2. Table reads ; In Grade 2, Book I, the average frequency of occur- rence of the Ayres-Teachers College words coumon to the textbook is 4.6; in Grade 3, Book I, the average frequency of occurrence of the Ayres-Teachers College words coiamon to the textbook is 4.1; etc. \ 99 - Book IX; and tenth, Book X. Average Repetitions o_f Ayres-Teachers College Words In the preceding discussion we were concerned with the average number of repetitions per word of all the different words in the vocabulary of each grade in each book. However, if the authors of the spellers were consciously providing for repetition of words, one is inclined to think that the "most frequently used words" would be repeated more than all words in general. Two causes would tend to bring about more frequent repetition of these words. First, their widespread popularization by means of the Ayres scale for measuring spelling ability and by the well-known results of the investigations of Cook and O'Shea, Eldridge, and others have brought them sharply to the attention of authors of spelling books. For some six or more years these words have not been a hidden secret in the science of education. In the second place, the spellers having dictation mater- ial would, by force of the necessity of sentence forming, if for no other reasons, have a greater average number of repetitions per word for the "most frequently used" words than would otherwise obtain. Therefore, we decided to find the average number of repetitions per word of the Ayres-Teachers College "most frequently used words" that are found in the various grades of the different spellers used in our study. The data so derived are shown in Table XXIV. In ° 9nera l we status for the average number of repeti- tions per word for the Ayres-Teachers College words common to the vocabularies of the grades as we did for all the words in the grades. That is, there is a larger average number of repetitions per word in the earlier grades than in the later grades. Again there is variation in the practice of the different authors- some having a relatively high average in a given grade and others a relatively low average excepting in the upper grades where the average is about the same in all the spellers. A careful comparison of the averages for the two sets will disclose J - 100 - that in general the averages obtained on the basis of the Ayres-Teachers College words common to the grades are slightly higher than the averages derived on the basis of all the words in the grades. The exceptions are Books IV and X. In Book IV, Grades 2, 3, 4, 5> and 6, the averages based on the Ayres-Teachers Col- lege words are slightly lower than the averages based on all words and in Grade 7 the two averages are the same. In Book X the two sets of averages are identical in four of the grades. Wo recognize that it is precarious to attempt to draw any conclusions on the basis of such slight differences, but the data show that Books IV and X are deficient in that they do not provide an appreciably higher average number of repetitions of the "most frequently used words." Repetition of Words from Grade to Grade It is a well-known fact that words in general present spelling diffi- culty in more than one grade. We are compelled to teach a given word to the children in more than one grade in order that they may learn it to the point of automatic reproduction. Hence, if a speller is properly constructed to take care of the teaching and reteaching of the words in its vocabulary, it must have its words appear in more than one grade. It would be an Herculean task to trace the entire vocabulary of a speller through the different grades, regardless of the fact that the work would be simple and straightforward. We were compelled to abandon such a task because j of lack of clerical help. However, we derived a quantitative expression for this j repetition from grade to grade by calculations on data derived chiefly for anoth- er purpose* Table XXV shows the original data and the calculations made in de- riving a general quantitative expression for the extent to which the words in the vocabulary of a given speller appears in more than one grade in the same speller. Owing to the difficulty of understanding the term "grade appearance" we shall give an illustration. Suppose that in Book IV the word "built" appears ■ v j ; o • j. • . 'i . ■, . • •. • . i i? r nl J - - 'i : ... * ■ .. . ; J ; . . J ... j ■ .-j. . . . : ... .. > i, : • • •* : 101 - TABLE XXV REPETITION OF WORDS FROM GRADE TO GRADE Numbei differenl • of b words Grade appear ance^^ Book Gross^ Net Amount^ Per cent ( c ) 1 2 3 4 I 5865 , 3968 1897 47.8 II 7052 5001 2051 43.0 III 7398 4787 2611 54.5 IV 7054 4742 2312 48.8 V 9549 5911 3638 61.6 VI 7731 5328 2403 45.1 VII 6654 4545 2109 46.4 VIII 9119 4767 4352 91.3 IX 8300 5915 2385 40.4 X 5221 4159 1062 25.5 { g ^ Entry for each book obtained by adding the number of different words in each grade. Entries in thi3 column are the differences between , . entries in Columns 1 and 2. ' c 'Entries in this column calculated from Columns 2 / , \ fiuid 3 • "Grade appearance" indicates that a word is found in a grade regardless of the number of times it appears in a grade after the first appearance. ~^ 19 o?f°r e ’ reads : In Book I there are 5365 printed words and 3968 different words of which 1897 were found in two or more grades. This comprised 47.8 percent of the total net vocabulary; etc. ♦ \ * •; - 102 - 3 times in Grade 3, 6 times in Grade 4, once in Grade 5, and once in Grade 6. In this case we would say that the word "built” made 4 "grade appearances," but in the net vocabulary for the book as a whole the word would appear only once. But in Table XXV the entry in column 3 would be only 3 because one of the 4 ap- pearances is found in column 2. The significant part of the table is column 4 which is concerned with the percent of the vocabulary of a speller that appears in two or more grades. Book X, having 25*5 percent of its vocabulary appearing in two or more grades, shows the least repetition from grade to grade, while Book VIII, having 91.3 percent appearing in two or more grades, has the largest repetition from grade to grade. This discrepancy is wide and represents well the fact that one author has constructed his speller so that repetition from grade to grade is an important factor while the other pays little attention to this principle. Approximately what percent of the vocabulary of a speller should appear in two or more grades we are not able to determine. However, if it is safe to argue from general prac- tice, we should say that 25*5 percent is too low because the next lowest percent for our ten spellers is 40.4 for Book IX, and that 91*3 percent is too high be- cause the next highest percent is 61.6 for Book V. However, if it is not safe to argue from general practice, we may conclude that both 25*5 percent and 91.3 percent of the vocabulary crossing grade lines cannot be the optimum status. One or the other or both are wrong from the standpoint of pupil needs in general. Since there is a possibility that dictation material might nullify to a certain extent comparisons between a speller having this form of presentation and one without it, this difficulty may be avoided by comparing Books VI and X both of which have practically no dictation. Book VI has 45. 1 percent of its vocabulary in two or more grades, while Book X has only 25.5 percent. Or, again, Book VIII has next to the lowest percent of dictation and by far the highest per- £ cent of repetition from grade to grade. This indicates that there is no very ♦ - 103 - close correlation between the proportion of dictation in a speller and the pro- portion of the vocabulary in the speller that appears in two or more grades* Therefore, we see wide range in regard to the application of the principle of repetition. ” Repetition of a Selected List of Words Since, as we indicated before, it was impossible for us to follow the behavior of each word in the vocabulary of a given speller through each grade in the speller, we decided to make this study for a selected list of words. Because of the wide popularity enjoyed by Jones's so-called "one hundred demons," it seemed likely that if any list of words was systematically repeated in a grade and in several grades of each of the ten spellers, it would be this list. On investigation we found that approximately all of the words of this list are found in all of our spellers. For purposee of comparing the repetition of these words with the diffi- culty of the 6ame words, we were compelled to drop ten of the Jones's "demons" and substitute in their stead ten other words from the Ayres spelling scale. We took the most frequent word from each of the last nine columns and from one col- umn we took two words. The following illustration for two words in one speller shows how the data were tabulated: Word Grade Total 2 3 4 5 6 7 8 1 blue 8 1 4 i 3 17 country 16 1 1 1 19 That is, the word "blue" in this speller occurred 8 times in Grade 2, once in Grade 3, 4 times in Grade 4, once in Grade 5, 3 times in Grade 6, and did not appear in Grades 7 and 8. Altogether it occurred 17 times in the speller* In order that we might obtain a quantitative expression for the » ' 'fa, C : ' 4 -"- • ■ . 4 - ' "x Y i 1 : . v 1 ■; . • . " . ; - •, fc O * t Off) i'. ; , , l) .... 0 ’ ... . . i. , " i •>■■■■ ■ ■ ..i ■ . .; w . < Y - 104 - repetition of these one hundred words, we determined their average number of re- petitions in each of the spellers in our study. The results of the calculations are shown in Table XXVI. TABLE XXVI AVERAGE NUMBER OF OCCURRENCES PER WCRD FOR SELECTED LIST OF 100 Book Average number of occur- rences per word Book Average number of occur- rences per word I 12.5 VI 7.3 II 6.8 VII 8.6 III 12.1 VIII 11.4 IV 8.8 IX 6.2 V 10.6 jlJ 2.5 Table reads ; In Book I the average num- ber of occurrences per word for the one hundred words is 12.5; in Book II the average number of occurrences for the one hundred words is 6.8; etc. Book I heads the list of spellers in the average number of occurrences of these well-known and popularly assumed one hundred "demons" of our language. ' Books III and VIII follow closely in second and third places. The lowest speller 1 in the list is Book X which provides for very little repetition of these words. Book VI may be more comparable with Book X than any of the other spellers in the list because neither of them has enough dictation material to influence mater- ially the repetition of the words. However, we find that Book VI has on the average 7.3 occurrences per word while Book X has only 2.5— but little more than one-third as many. In order that we may give the reader some idea of the detailed appear- ances of the "demons" in the grades of the ten spellers, we shall present in s Tables XXVII, XXVIII, and XXIX the behavior of three of these words. Lack of - 104a TABLE XXVII BEHAVIOR OF THE EASY WORD "SOME" Book Grade Total 2 3 4 5 6 7 8 I 12 9 6 3 3 33 II 4 4 4 3 3 18 III 11 14 8 13 6 52 IV 11 7 3 3 1 25 V 2 1 1 6 2 4 16 VI 4 1 5 VII 5 3 4 3 2 2 19 VIII 9 9 4 2 1 1 26 IX 3 2 1 4 1 11 X 2 1 4 7 Table reads : In Book I "some” appeared in Grade 2 twelve times; in Grade 3, nine times; in Grade 4, six times; etc., making a total of thirty-three appearances in the speller; etc. - 104b - TABLE XXVIII BEHAVIOR OF THE MEDIAN DIFFICULTY WCRD "WRITE" Book Grade Total 2 3 [_!_ 5 6 7 8 I 11 2 i 2 16 II 3 1 2 2 8 III 2 3 2 7 IV 2 4 2 2 1 11 V 2 2 2 1 7 VI 7 1 8 VII 1 2 2 3 2 10 VIII 3 7 1 3 1 15 IX 7 2 1 1 ll x 2 1 3 reads : In Book I "write” appeared in Grade 3 eleven times; in Grade 4, two times; in Grade 5, once; in Grade 8, two times, making a total of sixteen ap- pearances in the speller; etc. •104c. TABLE XXIX BEHAVIOR OF THE DIFFICULT WCRD "SINCERELY" appearances in the speller; etc. - 105 - space forbids our presenting more of them, regardless of the fact that the be- i havior of these "demons" provides much that is interesting. The easy word "some" (difficulty 26) is placed by all the ten spellers in the lowest grade in which they place any words. In general the word is re- peated many times in each grade and in several of the grades. The exceptions are Books VI and X. The range in total occurrences for this word is from 5 times in Book VI to 52 times in Book III. The large amount of dictation material in Book III doubtless accounts for its high frequency for that book. The word "write" which is a median difficulty word (difficulty 37) is placed first in Grade 2 by four of the spellers and first in Grade 3 by seven of them. It is found in more than one grade in all the spellers. The range in number of grades found is from two to five. The total number of appearances per speller ranges from three times in Book X to 16 times in Book I. That dictation material has little influence may be seen by comparing Book VI, having practical- ly no dictation, with Book III, having the largest percent of dictation. The total number of appearances in Book VI is 8 and in Book III, J. The word "sincerely" which is a difficult word (difficulty 54) has a more varied experience as far as the first grade introduction. In four of the spellers it appears first in Grade 5; in two of the spellers, in Grade 6; in i three of the spellers, in Grade 7; and in one speller in Grade 8. The total num- ber of appearances in the spellers ranges from once in Books II, IV, and IX to six times in Book VII. I The last paragraph indicates one of the reasons why the word "sincere- ly" is difficult to spell, that is, the children do not have enough experience with the word in their spellers. On the other hand, the word "some" may be easy oe cause the children have relatively many more experiences with it in their speller, as indicated by the fact that one speller prints it 52 times while the same speller prints the word "sincerely" only 4 times. This train of thoughts 8 t0 the furtller thought that perhaps many of our difficult words for | . . . i • . j . .j • . . . . . po . .. .. .. * . ‘ • *'• . V .. - . . . . ■ ■ .. - . : -.v . . ... . l .... y. ... i •. .'....1 ifiBCS aBB M • -L ' . . .. .■ . .' . . " . - - ' . '.- j *. Wffj . ; . ; . .j . ' . ..' . •: . . . , . V . • • '. . . .. t : J..! ■ ■ i . i ... « . . " •. : • ■» •. . . . ... . K - . ' ■ ' . .» “V ■ -0 2 •- *> . - 106 - children are difficult by virtue of the fact that they do not appear often enough in our spellers. The children do not have enough experiences with the difficult words to learn them. This brings us to a consideration of the question of the correlation between the repetitions in the spellers and the difficulty of the words in the spellers. This will be presented in the following sections. Correlation Between Repetition and Difficulty of Words At the close of the preceding section we suggested that, since the easy word "some” appeared so many times in the spellers and the difficult word Sin- cerely" so few times, perhaps this is one of the causes why our school children find the word "some" easy to spell and the word ’'sincerely' 1 difficult to spell. That is, the words are difficult to spell because in part, they are not presented often enough for the pupils to learn to spell them. Strong evidence supporting this position would be obtained if we should find upon investigation that there is a negative correlation or no correlation between the repetition and the diffi- culty of the words. We were not able to make the preceding study for all the words in the spellers, because we do not at present know the difficulty of all the different words of the vocabularies of the ten spellers. We were, therefore, limited in the scope of our investigation to words of known difficulty. In other words, we j were compelled to confine our study to the words common to the Ayres-Teachers College list of "most frequently used words” and to the spellers. When the vocabulary for each speller was alphabetized for each book as a whole, we constructed a correlation table for the frequency of occurrence and for the difficulty of the words common to each speller and to the Ayres-Teachers College list. This gave us data for approximately 2000 words for each speller. The values for r calculated by the Pearson product moment formula are shown in Table XXX. 5 - 107 TABLE XXX CORRELATIONS BETWEEN REPETITION AND DIFFICULTY BASED ON WORDS COMMON TO EACH SPELLER AND TO THE AYRES-TEACHERS COLLEGE LIST Book r Book r I - 0.26 VI + 0.20 II - 0.29 VII - 0.26 III - 0.30 VIII - 0.18 IV - 0.32 IX - 0.23 [ V - 0.22 X - 0.07 Table reads : la Book I the correlation between the repetition of the words common to it and the Ayres-Teachers College list and the dif- ficulty of these words is - 0.26; in Book II the correlation between the repetition of the words common to it and to the Ayres- Teachers College list and the difficulty of these words is - 0.29; etc. The table shows that there is an appreciably large negative correlation between the repetition of the words common to the Ayres-Teachers College list and to the spellers and the difficulty of these words. The two exceptions are Books VI and X. Book X has such a small negative correlation coefficient that we may say that there is no correlation between the repetition and the difficulty of the words common to it and to the Ayres-Teachers College list. However, for Book VI the positive correlation is about as large as the negative correlations for eight of the other spellers. These data show that in eight of the spellers the easier words occur more frequently than the more difficult words. Hence, the pupils have more ex- perience with the easy words which aids in keeping them easy and less experience with the difficult words which aids in keeping them difficult. In studying repetition of words we need to keep in mind that some of the words in the English language are needed in sentence forming. They appear frequently in the spellers having dictation because of this necessity. We - 108 - excluded, as indicated in Chapter II, twenty- four of these easy words which con- stitute over 27 percent of running correspondence material. This makes such sentence-forming words have less influence on our coefficient of correlation between repetition and difficulty. Since the words in the Ayres component occur more frequently than the words in the Teachers College component, the chances are favorable for the Ayres words to appear more frequently than the Teachers College words in the spellers having dictation. That is, Teachers College words, it woul< seem, have a better chance of appearing in spellers by design on the part of the authors than by chance sentence- forming needs. Therefore, we calculated the cor- relation between repetition and difficulty of the words that are common to each speller and to the Teachers College list of "second and third thousand most fre- quently used words.” The values of £ obtained from these sources are shown in Table XXXI. TABLE XXXI CORRELATIONS BETWEEN REPETITION AND DIFFICULTY BASED ON WORDS COMMON TO EACH SPELLER AND TO THE TEACHERS COLLEGE LIST Book r Book r I - 0.14 VI + 0.21 II - 0.09 VII — 0.12 III - 0.32 VIII - 0.12 IV - 0.35 IX - 0.19 V I - 0.57 X + 0.32 Table reads : In Book I the correlation between the repetition of the words common to it and to the Teachers College list and the diffi- culty of these words is - 0.14; in Book II the correlation between the repetition of the words common to it and to the Teachers College list and the difficulty of these words is - 0.09; etc. As in Table XXX we find appreciably large negative correlation between tv. .,-2 ' i 1 1» ' . - 109 - table XXXII CORRELATION BETWEEN THE REPETITION OF THE WORDS IN THE SPELLER AND THEIR DIFFICULTY. (The value of £ was derived from only the word6 common to the speller and to the Ayres-Teacher s College List.) 1 Grade Book I II III IV V VI VII VIII IX X 2 - o . o 6 - 0.14 - 0.49 - o . o 8 (a) ( a ) - 0.25 - 0.16 ( a ) 0.02 3 - 0.01 - 0.24 - 0.14 0.02 - 0.01 0.22 - 0.02 0.17 - 0.12 0.09 4 0.02 - 0.37 - 0.11 - 0.06 - 0.12 0.30 - 0.21 0.05 0.05 0.001 5 - 0.02 0.21 - 0.21 0.05 - 0.15 0.03 - 0.03 - 0.02 - 0.17 0.18 6 0.02 - 0.17 - 0.17 0.45 - 0.02 0.04 - 0.20 0.02 - 0.17 0.03 7 - 0.04 - 0.11 0.22 0.08 - 0.12 0.40 - 0.22 i o o —5 - 0.20 0.03 8 - 0.02 - 0.24 0.01 - o . o 6 - 0.27 0.19 - 0.11 - 0.11 - 0.07 - 0.18 These hooks do not provide for spelling in Grade 2. Table rgads: In Grade 2, Book I, the correlation between the repetition of the words in this speller that are common to the Ayres-Teachers Col- lege List and the difficulty of these words is -0.06; in Grade 3, Book I, the correlation between the repetition of the words in this speller that are common to the Ayres-Teachers College List and the difficulty of these words is -0.01; etc. - 110 - the repetition and the difficulty of words in eight of the spellers. However, in this table we find two appreciably large positive correlations — Books VI and X* It will be noted that Book VI has approximately the same value for r in both tables, but Book X shows a strong positive correlation in this table, whereas in the former it showed practically no correlation. However, the large positive value of r means little for Book X because the average repetition is less than 2, as shown in the earlier part of this chapter. Correlation Between Repetition and Difficulty b£ Grades in Each Book In the preceding discussion of the correlation between difficulty and repetition, we were concerned with the spellers as a whole. In our search to determine the true status of the spellers in this regard we examined them grade by grade. We calculated the value of r on the basis of the Ayres-Teachers Col- lege list of "most frequently used words” and on the basis of one of its compo- nents— -the Teachers College "second and third thousand most frequently used words’! As mentioned in the preceding section, we made this second series of calculations in an effort to avoid certain words because "they are true construction words and necessary, no matter what the nature of the subject under consideration." The correlation between the repetition and the difficulty of the words common to each grade in each speller and to the Ayres-Teachers College list is shown in Table XXXII. I Because of the prevalence of negative signs and the small values of r , the general impression from the table is that the correlation between repeti- tion and difficulty of the words in the grades is either inverse or absent. An examination of the size of the positive values of r, (e. g. consider the positive values of r that are smaller than 0.1C) further confirms the impression that there is either a negative or zero correlation between the difficulty and the repetition of the words. To be sure, this is assuming for all words in each grade the same status as found for the words common to each grade and the Ayres- 1 ^ J. ' 1 • . • . . * i J . - - • . - , / • . ■ ■ v r _ -I Q-J ; ' ' . .. . ' ' . : •- 'i ■? •- i. . • * ‘ • • . ■ . . • ‘ l - Ill - 112 - Teachers College list. On the other hand, when we exclude the Ayres words from consideration, the values of r based on the Teachers College list alone (Table XXXIII) show a preponderance of positive correlations for the grades. In fact, there are 53 positive correlations and only 14 negative ones. But only a few of the positive correlations are large enough to be significant. None is as large as 0.40. The average of the positive correlations is 0.175* Thirteen of the correlations are less than 0.10. We are pretty well agreed that such coefficients have little meaning. It is difficult to say what value of r would be significant. However, for all values below 0.25 we are probably safe in assuming that the correlation is not significantly large. Therefore, comparatively few grades have a signifi- cantly large value for jr. Only 13 of the 67 grades in the ten spellers meet this assumed standard. The reader will also observe that the negative and extremely small positive correlations tend to appear in the lower grades. For example, 9 of the 14 negative coefficients are attached to Grades 2 and 3, and no higher grade has more than one such coefficient. Two considerations lead us to believe that authors are more successful with hard than with easy words in making repetition H go with'* difficulty. The firet consideration is that a preponderance of posi- tive coefficients only appears when we consider the Teachers College list alone— i* e., the hardest list. The second is that even with that list the authors are most successful in the upper grades where naturally the words are the hardest. The correlations between the repetition and the difficulty of the words common to the different grades and to either the Ayres-Teachers College list or the Teachers College list alone are either negative, none, or not large enough to be of real significance. All of which confirms the statement that our modern spellers are propagating correct spelling of the easy words which will therefore always remain easy, while the present difficult words will always - 113 - remain difficult because of insufficient repetition to make them relatively easier. At this point it is to be regretted that we do not have reliable data on the learning difficulty of words which may not be the same as the spelling .difficulty. If we had the learning difficulty expressed in terms of the factors of habit- formation— especially repetition— cur comparisons could be made direct instead of indirect as ours have been, i. o., by assuming that one of the reasons why words are relatively difficult to spell is because of the relatively few experiences that the school children have with spelling these words. We might come to the same conclusions about the spellers that we have just reached; but the argument would be more direct and perhaps more convincing. Therefore, we suggest that one of the problems for future research is the derivation of a learning difficulty factor for each of the three thousand words in the Ayres- Teachers College list and that this factor be expressed at least in the number of repetitions necessary to insure permanently correct recall. From the viewpoint of the teacher these data on correlations between difficulty and repetition suggest that he should know the spelling difficulty of the words taught and that he take measures to provide for many more pupil exper- iences with the difficult words than are offered by the speller. From the view- | point of the construction of spellers, provisions should be made for much more repetition of the relatively more difficult words. Further, the speller would immeasurably aid the teacher if certain devices were used to indicate the diffi- culty of the words. This might be done in an alphabetical "finding" vocabulary for each grade. Finally, from the viewpoint of the authorities who select the spellers, they should know, among other things, what provisions are made for repetition of the words according to their difficulty. Correlation Between Repetition in Spellers and in Adult Usage In the preceding section the data indicated, in general, either no - 114 - correlation or negative correlation between repetition and difficulty of the words in spellers. That ie, in the first case there was no relation between the spelling difficulty of the words and the number of times they appeared in the spellers, and in the other case the easy words were repeated more than the diffi- cult words. Therefore, we assume that the principle of difficulty did not func- tion in the arrangement of the words because, if it had, the more difficult words in the selected vocabulary would have been repeated more than the easy ones. However, it is possible that the well-known principle of use in curric- lum construction might have been the criterion of the authors of the spellers. If this principle has functioned, we should expect to find a high positive cor- relation between the repetition of words in the spellers and the words used in written work outside the spellers. The only words for which we have a frequency factor are those derived from adult writing-correspondence, newspapers, etc. Hence, the phrase ’’written work outside the spellers" must be confined to adult usage as stated. One would expect to find a high positive correlation between repetition 5 in the spellers and in adult usage for three reasons. First, the authors of the spellers, as mentioned in Chapter I, state in their prefaces that they had made use of recent scientific studies of spelling vocabularies and most of these were based on adult writing in which frequency of use was the criterion. Second, the prevalence of dictation material would tend to bring about repetition of words, especially those necessary in writing about any subject. Third, absence of positive correlation between difficulty and repetition, as shown in the foregoing section, might lead one to think that the principle of use is functioning because "to a certain extent there is an inverse relationship between difficulty and frequency of use."^ 3 Buckingham, B. R. "Scientific curriculum building," (Editorial), Journal of Educational Resear ch, 1:406, May, 1920. ( • ' • : ■ ■ J Cl j *1 i; _ s»| . V. . . j . >J . ■. ■ -C ■j - : * ■ * V . c. ■ - i •**■9 • - ' ’ . '• - • • / t .. . . . * • . > ■* .. l . . . . . . .. .■ • « 'i ^ ; , . 1 . - . J •' O'- - . ,.'i ' ! j . ... > - 1 ■ ;; .* ■ • - - .... ' * *- .J- ji -' . . • • .... i *' •J V ■ f J 115 The principle of use may operate in curriculum construction in spelling in one or both of two ways. First, we may place in the spelling vocabulary words within certain specified limits of frequency of occurrence in their H natural set- ting." Second, we may place the words as frequently in the spellers as they occur in their "natural setting." When this is done, the first case is included. We have covered the first case in Chapter IV and we shall deal with the second in this chapter. In order that we might test the extent to which there is correlation between the frequency of occurrence of words in spellers and the frequency of occurrence of these words in adult usage, we were again limited to the words common to the spellers and to the Ayres-Teachers College list because these words have a known frequency. In the second place, we found it impossible to calculate the value of r on all the words as derived by the first limitation because of lack of time and clerical help. Therefore, we made a random sampling of words-- 50 from the Ayres component and 50 from the Teachers College component— and cal- culated the correlation between the frequency of occurrence of these words in the spellers and in adult usage. The values of r thus derived are shown in Table XXXIV. TABLE XXXIV CORRELATIONS BETWEEN REPETITION IN SPELLER AND IN LIFE BASED ON 100 WORDS SELECTED AT RANDOM FROM THE AYRES-TEACHERS COLLEGE LIST AND COMMON TO EACH OF THE SPELLERS Book r Book r I 0.004 VI 0.76 II 0.79 VII 0.79 III 0.49 VIII 0.83 IV -0.02 IX 0.72 V 0.90 X 0.66 116 Table reads : In Book I the correlation between the frequency of occurrence of 100 words se- lected at random from the Ayres-Teacher s College list and the frequency of occurrence of these words is 0.004; etc. Looking over the table we see that eight of the ten spellers have a positive correlation between repetition in the spellers and in adult usage, while two of the spellers, Books I and IV, show practically no correlation. Little relation if any exists between the proportionate amount of dictation because the eight spellers having the high values of r have both the highest and the lowest proportionate amount of dictation, as may be seen by comparing the data in this section with the data in Chapter VII. As stated in the introductory paragraphs of this section, there should be an inverse relation between the principles of difficulty and of use. That is, the larger positive values of r in this section should be paired with the larger negative values of r in the preceding section. However, a comparison of the corresponding values of the r^s in the two tables shows that no such close inverse relation exists. There is another sense in which we might interpret inverse relation, i. a. , the higher the correlation between the repetition in the spellers and the repetition in life, the nearer the correlation between difficulty and repetition in the spellers approaches zero or no correlation. However, again, thi 3 relation does not exist as may be seen by comparing the values of r in this section and in the preceding one. Therefore, the principle of use as we have interpreted it does not seem to take the place of the principle of difficulty or the opposite. The one speller that attracts our attention is Book VI. This speller exemplifies a combination of the principles of difficulty and of use in curric- ulum construction that is highly desired. That is, there is a relatively high correlation between repetition in the speller and difficulty of the words, and a high correlation between repetition in the speller and repetition in life or adult usage. . * . J • \ J . ■- . - . . . •i • -■ • ' . • .. U ■ 1 /J . • - n! , . . v • : . ’ ' • . , - . •. - ... . .■ ■ . u • . • • . 7 u ■ ‘ . • . . .. . . • •: : ••• • O' ' ' . - 117 - Some suggestions concerning review of words may not be out of place. High frequency words in life should be in the spellers. The difficult words that are frequently used should have high repetitions. Of two words each equally common, that word should be repeated the more which is the more difficult. Sim- ilarly, of two words of equal dif ficulty— that difficulty being relatively high— the more common word should be repeated the more often. However, if the two words, having equal difficulty, are easy, there appears to be no reason why the more common one should occur more frequently in the spellers. This is because the repetition in life will take care of the spelling of such words. - 118 - CHAPTER VII FORMS OF PRESENTATION OF WORDS The textbook plays a very important role in American education. To a large majority of our elementary teachers it furnishes all the subject matter taught to the pupils and the chief guide to the methods of teaching. However, it should be noted in passing that this is not so much an indictment of the teachers as it is a suggestion to the authors of textbooks that they select the contents as carefully as possible in the light of scientific investigations and use the best established forms of presentation for this content. The significance of the forms of presentation lies in the suggestions to teachers of the methods of teaching that may be employed in assisting pupils to gain control of the subject matter. For example, if the adopted speller pre- sents the words in column form only, the teacher will be more likely than would be the case with a different form of presentation to have the pupils confine their spelling work— both study and reproduction— to column form. The pupils will be required to study the words in column form, spell them orally to the ! teacher as h© pronounces them from the columns, and to write them in column form whenever the words are written for purposes of the spelling lesson. On the other n&nd, if words are presented in the speller in both column and dictation forms, the pupils will study the words in both settings, and the teacher is more likely to have the pupils write them in column and in dictation forms as methods of reproduction, because of the example set in the speller. j One of the purposes in our project in the analysis and measurement of spellers was to express quantitatively some of the more important forms of word presentation. As mentioned in Chapter II, the clerks as they transferred each word from the speller to the cards entered thereon a symbol indicating the form of presentation in which the word was found— whether in columns, in column review, in supplementary lists, in dictation, in dictation review, etc. All the data obtained for this chapter were taken from the master cards. If we wished to determine the number of column words, we counted the C’s (the symbol for column words) on the master cards; or if the number of column review words, we counted the number of CR*s (the symbol for column review words) on the master cards; or if the number of dictation words, we counted the number of D»8 (the symbol for dictation words) on the master cards; etc. The foregoing method was used while the master cards were still in alphabetical arrangement b£ grad es in each speller; and in order to find the total for a speller for a given item, 0» g« C!» the grade items of the same kind were added. The reader should be cautioned to consider the data in this chapter more indicative than absolute, because of the many difficulties encountered in J this phase of the project. First, as implied in Chapter II, it was not always possible to decide to our entire satisfaction just the category in which to place a given set of words. This confusion was due in many cases to a different nomenclature used in the spellers, and in many cases to the different contexts in which the words were set. Second, errors no doubt slipped into the work because of the necessity of training the clerks on the job. Third, the handling of thousands of cards made it impossible to prevent some errors from creeping into the work. However, we exercised all the care possible and we feel reasonably sure that we have not made very large errors in the counts for any of the differ- ent forms of presentation. Another limitation to the investigation of the forms of presentation is the lack of conclusive experimental data on the effects of the different forms of presentation. For example, we do not know whether column form of presentation alone is superior to or inferior to, say , dictation form alone; or whether a com- bination of column and dictation forms is superior or inferior to either the one . i r Eu . . . . . . . . . > . * • - . ' ’ ■ ... .. - ’• » S : 1 . .j: . . , ■ - 120 - or the other alone. Consequently, we shall have to wait on the development of experimental pedagogy before we can say conclusively that one speller is better than another because of a given form of presentation in one and not in another. Therefore, we shall have to content ourselves with exhibiting the findings, show- ing general practices, and variations in the practice for a given form of pre- sentation. Percent of Spellers Consisting of Dictation Material With the foregoing statements of the limitations in mind, we shall proceed to exhibit the findings in the ten spellers. First, we shall call the reader's attention to the percent that dictation material^" is of all printed 2 material in the spellers. The mass picture for each of the ten spellers is shown in Table XXXV. TABLE XXXV PERCENT OF SPELLERS CONSISTING OF DICTATION MATERIAL Book ipercent of speller in dictation Book Percent of speller in dictation I 35*5 VI (b) II 51.3 VII 35.4 III 65-5 ■1J i . , ' . . • '■ • . ' , , - . .. • J . ■ ‘ V ' 3 • //'yure 8 Perce/?/ of 8pe//ers Ca/ztposec/ of D/c fo f/ 0/7 Ma fer/a/ //? 7yp/ca/ Books Grade * - ■- 1 . ' 1 -Jr _± TABLE XXXVIII PERCENT 0, CERTAIN FORMS OF PRESENTATION IS OF ALL FORMS OF PRESENTATION Grade Percent Column Form is of all forms Percent Column Re- view Form is of all forms Percent Dictation Form is of all forms Total the three forms 2 30.5 17.4 38.3 fc= ct ■ B1 86.2 3 33.7 18.0 36.7 88.4 4 33.2 16.8 36.6 86.6 5 35.5 17.3 39.8 92.6 6 37.5 18.4 37.4 93.3 7 42.3 22.0 34.4 98.7 8 41.8 19.5 31.4 92.7 Calculated on number for each iters concerned and on the total number for all items. Table reads : In Grade 2 column form of presentation is 30*5 percent of all forms, column review form is 17.4 percent of all forms, dictation form is 38.3 percent of all forms; and the total of these three forms is 86.2 percent of all forms; etc. Consulting Table XXXVIII one sees that present practice, as determined by our spellers, decrees that the proportion of column form of presentation shall increase with advancing grades, that the column review form shall remain about on a level, and that the dictation form shall decrease slightly. Further, he will observe from the column which shows the percent these three forms of pre- sentation is of all forms of presentation that a very large proportion of the spellers is devoted to these three forms. For example, in Grade 7 we see that forms of presentation other than these three practically disappear. Collation of All Forms of Presentation Examined In the preceding section we exhibited certain of the more prevalent forms of presentation of words. However, for purposes of giving the reader a i - 127 - complete display of the different forms and in order that we might make further detailed comparisons, we are giving in this section a complete collation of all the forms that were examined in our project. Number of printed words . --The number of printed words in each form of presentation for the ten spellers used in our study is shown in Table XXXIX. These figures are taken from tables which are not shown in this dissertation, but which are on file in the Bureau of Educational Research, College of Education, University of Illinois. A study of the ranges in the number of occurrences of words in the dif- ferent categories reveals interesting information. The range in the number of column words is from 3903 in Book X to 7009 in Book VI. The average of the ten books is 4995 column words. The range in the column review words is from 540 in Book X to 703? in Book VI. The average is 2516 column review words. The next form having entries for all books is dictation. The range is from 315 in Book VI to 11,189 in Book III. The average is 4225 dictation words. In a study of the number of words in dictation it should be recalled that 24 small easy words aggregating approximately 27 percent of all the running words in correspondence were omitted. Hence, in getting a true picture of the amount of dictation, these 24 words with their percent of frequency should be taken into account. The range of the number of words in word-building form is from xero in Books III and X to 707 in Book IX. The average is 404 on the basis of the eight books having this entry. The range for word-analysis is from zero in Books I and X to 392 in Book IX. The average of the eight books is 125» The range for words in phonic form of presentation is from zero in Books I and X to 1132 in Book VI. The average of the eight books is 488 words in phonic form. The other forms show scattering entries for the ten books and comparisons are not made. The total heading at the right side of Table XXXIX gives an idea of the number of printed words. The range is from 5^96 printed words in Book X to — — — ' - * ' . 1 ; : -V . . . '■ . i . • . . * *y t • ■ * . • , . ■ . ' . V • ■J - • . \ . . : - K . . • ■ ' . ■ •' : ■- ■ - 128 - In Book I there are 54-71 words printed in columns, 2103 words printed i column review, 1591 words printed for supplementary work, etc. 129 - to 19,523 in Book III. The average of the ten books is 13,623. One cause of the large number of printed words in Book III is the great amount of dictation and one of the causes of the small number of words in Book X is the fact that it has no dictation. In comparing Books V, VI, and IX with the others it should be kept in mind that these books do not offer spelling for Grade 2 while the others do. Books V and VI are above the average in number of printed words regardless of the fact that they have one less grade. It should also be noted that Book VI has a negligible amount of dictation. Hence, it does not always follow that the spell- er having the greatest amount of dictation will have the largest number of printed words. To summarize, the data show wide variation in practice in the number of words appearing in a given form; the total number of printed words in a text does not have perfeot correlation with the number of words in dictation; and there is no general agreement as to the different forms of presentation that should be used. These facte suggeet that the field of methods of teaching spelling ie for- tile and fallows j£?cetttage each form of presenta tion is of all forms in the spellers . — For purposes of comparison of the different forms of presentation in each speller the percentage system has certain advantages. Table XL, which was derived from Table XXXIX, shows the percent that the number of printed words in each form of presentation is of all the printed words in the speller. A survey of Table XL shows that all the spellers present words in col- umns, in column review, and in dictation. Column review ie used to a very small extent in Books II and X. Dictation is used to a very small extent in Books VI ana IX. Further, one notes that these three methods form the major portion of the total different occurrences. Over three-fourths of the printed words are found in these forms of presentation, if all the books are considered as a unit. . *-» . . 4 ' . : - •“ • ■ t- ■ * . “ - > - • * ' . _ • ' . . . ■ ■ ' • ' , , TABLE XL PERCENT THAT THE NUMBER OF PRINTED WORDS IN EACH FORM OF PRESENTATION IS OF ALL THE PRINTED WORDS IN THE BOOK - 130 - ra E u o -p © 5 - © •H T) © 43 -P © -P rf o •H •n c •H O -P © ro 3 o a >» © © 43 -p «M o © c o •H -P •H S3 •H V. © T3 © .C -P s* • O C l M O M -P M aS -P Li S3 © © -p © G, © © L, 42 Pi O Vi © O © to © Table read s; In Book I the words printed in colurans comprise 35«3 percent of all the printed words, the words printed in column review comprise 13.6 percent of all the printed words, the words printed in supplementary work comprise 10.3 percent of all the printed words, etc. Such forms of presentation as word building, word analysis, and plural formations take up a very small percent of the books. These methods of rationalizing spell- ing are not in favor with the textbook makers of spellers. The layman will no doubt be surprised to find that such a small percent of the pupil's time for spelling is taken up with such rationalizing processes. No one doubts that the outcome of instruction in spelling is a clear image of the words taught, but it is seriously open to question whether major reliance should be placed upon the presentation of the words in columns, in column review, and in dictation. How- ever, this is a question that will have to be determined by extended experiment. One observes that many of the texts make a slight use of many of these forms of presentation. Book X, for example, confines its forms to the first four listed in Table XL. The reader will be interested in knowing that in this book not more than 80 words a grade appear in column review form, that an average of less than 100 words a grade appear in supplementary form, and that the dictation appears only in four letters at the end of the work for Grade 8. Further, it is observed that no one speller uses all the different forms of presentation. There is a wide range of practice among the ten spellers as to the percent of space allotted to a given form of presentation. Under the column form we find that the range is from 20.9 percent in Book III to 66.0 percent in Book X. Column review ranges from 5.8 percent in Book II to 42.4 percent in Book VI. Dictation ranges from 1.9 percent in Book VI to 57.4 percent in Book III. The dictation in Book VI appears in Grade 4 only while the percent of dictation in Book III would perhaps have been higher if the long literary selections in the upper grades had been included in the study. In the three major ways of presentation of words the spellers are decidedly lacking in uniformity. The present lack of scientifically determined standards of methods of teaching spell- ing are vividly pictured in these ten spellers. The whole field of methods of teaching spelling, as indicated before, is wide open and there has been very little work done on these problems. . . ' . J . • v . . ' . • v - :. .. . .s . ' .... . ♦: 'X ' •.»’! rfJ • ;J/i* • f. ' ... - • .;.v. • . . ■ . ■ . • - -• . - 132 CHAPTER VIII TTO DIMENSIONAL CHECK LISTS The reader has doubtless wondered how superintendents or textbook com- missions might devise and use a shorter method for obtaining similar data on spellers under consideration for adoption because such elaborate and extensive methods as described in the preceding paragraphs are not feasible for these school people. In order to answer an inquiry of this nature we devised a short method of comparison of the selection, gradation, and repetition of the word offerings in the grades. This method involves the use of a check list of words, the object be- ing to find out how many of these words appear in the spelling book which is un- der examination. We have called our check list *two dimensional* , because its words exhibit two aspects or dimensions, namely, usefulness and difficulty. The bases of our check lists are Anderson’s^ three thousand words and 2 Ashbaugh's Iowa Spelling Scales. The first step in the derivation of the check lists was to assign Ander- son’s three thousand words to Grades 2 to 8 inclusive. It should be noted that Anderson’s list is arranged in descending order of frequency of occurrence of the words. Therefore, the first 180 words on the Anderson list were assigned to Grade 2 on the assumptions: (a) that the most frequently occurring words should be learned by the lowest grade; and (b) that this grade could be reasonably ex- Anderson. The determination of a spelling vocabulary based upon written correspondence . University of Iowa7 1917. 2 Ashbaugh. Iowa spelling scale s, University of Iowa, 1919. 'i •; ■ . . • ’ •- v: r ' ... . •* . v : -a . J. . , , • > \ ■ Kj. ife'ii a ■ ■i * v). ‘ i ’ . • ‘ . \. .... ■; c ‘is. ; .... .-.H . t-< . ‘ j v.; j .. .- ' * , ' “ . : . ; . i t •■■■>■ . : ' ■-.< ‘I .; . ....... ... * 133 - pected to learn one new word a day for l 8 o days of a school year. The next 360 successive words were assigned to Grade 3 on similar assumptions except that this ^rade might he expected to learn two new words a day for 180 days. The remain- der of the Anderson list was assigned in a similar manner to Grades 4 , 5 , 6 , 7, and 8 . The numbers of successive words were assigned as follows: Grade "Number 0 r Successive Words Gr ade ~ *%um5eT 'oT m Successive Words 2 180 5 360 3 n 3 n 360 6 540 4 360 7 540 8 7 64 a Tne assignment to Grade 8 is irregular in number because to it was apportioned all the words remain- ing in the Anderson list after the assignments were made to the preceding grades. These words with the Anderson frequency and the Ashbaugh spelling percent of ac- curacy found in the Iowa Spelling Scales were written on 3” by 5" cards and alphabetized within each grade. Tne words for each grade as determined by the preceding method were subjected to still further limitations by reference to their spelling percent of accuracy as shown by the Iowa Spelling Scales. That is, from each grade list were excluded those words whose percent of spelling accuracy as shown on the Iowa Spelling scales for the given grade fell between 73 and 100, and those words whose percent of spelling accuracy fell between 27 and 0 for the given grade. Tne words above 73 percent accuracy were considered too easy and those below 27 percent accuracy were considered too difficult. The limitations laentioned in the preceding paragraph reduced the words as selected from the Anderson list to 120 for Grade 2, 229 for Grade 3, 197 for Grade 4, 156 for Grade 5, 135 for Grade 6 , 102 for Grade 7, and 83 for Grade 8 . It will be noted that this limitation left more than 100 words for each grade except Grade 8, In order to get 100 words for this grade seventeen were taken from Buckingham’s Extension of the Ayres Spelling Scale. These seventeen were within the before mentioned percent of spelling accuracy limits used for the other grades. Then by random sampling except in Grade 8» 100 words were chosen. The lists thus selected constituted our two dimensional test lists for the grades These lists are reproduced in full at the end of this chapter. In the preceding paragraph we have sketched the method by which we de- rived a checking device that may be used in comparing the selection, gradation, and repetition of the grade vocabularies of spellers. To be sure, it is only an alternative for a more complete analysis of the contents of spellers. However, superintendents are often compelled to use short methods in order to get busi- ness done. Even this method will not be short unless each textbook provides a finding list of all the words contained in it. This reminds us that textbook makers should protect themselves against inadequate measures of the extent, sel- ection, gradation and repetition of the contents of their spellers by furnishing a complete alphabetical list of the word offerings for each grade. We had, however, the vocabulary of each one of the ten spellers, and in Table XLI we are reporting the results of the application of our two dimen- sional check list to these spellers. Looking at the totals near the bottom of Table XLI one observes that Book X has the greatest number of words common to the lists and to the grades— 240 words in all, while Book II has the least number common to the lists, and to the Grades — 96 words in all. On the basis of the number of words common to the books and the test lists, which measures selection and gradation of the vocabu- lary the spellers rank as follows: J • ' - . * ' < *: t ... >; , • • j.X • ■ ■ ■ c • . . — \ • i ' . ■... * - • i : TABLE XL I DATA OBTAINED BY THE SERIES OF WO DIMENSIONAL CHECK LISTS - 135 - 1 X eeouej -jnooo 1 o 1 oo 1 CO CO CO CQ rH CM CM CM CJN vr\ o OO CO CM \ i. H o o m epjo/iv JO jreqwmM o c— xr\ CQ CQ CQ Os rH CM CM ON CM c— o M CM j i I ! i x M eeouej j -jnooo © CO xr\ CO xO pH ON rH rH oo CM NO CO pH 1 t* j ♦f f vO O o CQ j epjOM JO J jeqranH © nO CO CQ CM CM rH Xr\ rH rH rH rH rH rH • ! rH ! M M H > ] eeouej -jnooo > (Tv ! xO 1 CM **fr t— rH ir\ its ON CM CQ lT\ H- rH vr\ c- o H- ! i 1 l ©> ] M O o i « SpjOAV jo jequmN C*— ITS OO pH CQ CQ. rH CQ CQ ON CM CQ CQ t- OO rH 2 . 1 * § I j M 1 Hi 5 > eeouej ; -jnooo rH pH CM CM vO ON CQ ON CM -M- CM C— CQ CM CM CQ pH CM I j NO j 2 o a O 5 ' 111 * J epjOAv JO jQqmnN CM CQ rH CM O CM O CM CM CM c— rH CM 1 CQ 1 pH H > eeouej -jnooo is O OO rH CQ ON ON CQ vO Os CQ vO 1 OO OO t • | o s '■** S CQ epjojft jo jeqtnnN i3 t— IT\ O CQ CQ rH CM CM rH NO CM CO 1 SD 1 rH s CM • CQ 1 > eeouej -jnooo i aj CM O rH CM CM o CQ CQ i ! oo j © o CQ 1 8pjCM jo jeqmnN ^0? OS Vr\ CM rH CM CM ON CM r— rH oo pH • l ! > i M ; eeouej -jnooQ] *H sD pH cr pH rH CO ••sf CQ CM o CQ CM iH CQ ’ ON e— CM O o 2 CQ J epjOM jo jeqranN CO rH CM rH rH CM CM vO rH o CM CQ CQ pH • O u Q, •P O s o TJ ro o © 43 © © © © © © 45 g © **-* TJ © -P -P a o CM © TJ © U es -p © *ri -P © © -P 73 c o •H W c © S •HI TJ O ?S -P © 45 55 •H © . TJ O fc. -P O © 5s O ©" o © - 1 J © -p 45 -p c— xO CM O tj co © rt -P J* T3 -H $5 t, 3 a, © <*-< © © © fc. £ p ^ m tj ► JU M s © CO O ■'* CQ © - © CM © 45 © -P Tj © TJ d § C © M TJ © hO ©J45 J-f-p ©J fc. rHI O rO[«M 5j{ H - 136 - Rank Book Rank Book 1 X 6 IV 2 VIII 7 VII 3 VI 8 IX 4 III 9 I 5 V 10 II However, the number of words which permits comparison of selection and gradation is not the only criterion by which we may compare the spellers on the data in Table XLI. Another factor may be considered, namely, the number of oc- currences for the words of the list that are in the spellers. These occurrences, as indicated before, provide for the automatic operation of one of the laws of habit -formation (exercise) in so far as the maker of the textbook can provide it. The average number of occurrences per word in these check lists is shown for each book at the bottom of Table XLI. In comparing the spellers we are rank- ing the book first that has the highest average number of occurrences. The ranks of the ten spellers are as follows: Rank Book Rank Bo ok 1 I 6 V 2 VI 7 II 3 III 8.5 VII 4 VIII 8.5 IX 5 IV 10 X It is interesting to compare these rankings on the basis of occurrences with the rankings on the basis of number of words common to the lists and the textbooks. It may be noted that Book I ranks first on ths basis of occurrences and ninth on the basis of the number of words common to the book and to the - ' ■ t . . , : • - ‘ 1 ( . . . ' . - 137 - 1 istg, while it may be seen that Book X ranks first on the number of words com- mon to it and the lists and tenth on the basis of the average number of occur- rences per word. i->ei us assume that Book I is excellent from the standpoint of occur- rences and one of the poorest from the standpoint of word selection and grada- tion, and that the reverse is true of Book X. Then the question arises as to which is the better speller when these two standards are both considered. Since we know of no accurate way of weighting the two criteria other than allowing equal weight to them, we shall add the ranks on the occurrences and on the sel- ection and gradation of words. The book having the smallest rank sum we shall give a new rank of on^, the book having the next to the smallest rank sum will be called number two, etc. The results are as follows: 1 Rank Book j Rank Book 1 VI 6 V 2 VIII 6 X 3 III 8 VII 4 I 9 IX 6 IV 10 II On the basis of these two criteria, Book VI stands at the head of the list and Book II at the foot. The claim may be made that this method is not the proper one by which to compare the spellers but it appeals tc us as more quantitative and impersonal than any other methods we have found. The foregoing is an illustration of what any superintendent and his corps of teachers may do in comparing spellers as to their selection, gradation, and repetition of words. This is a short and simple task for the spellers that provide an alphabetical finding vocabulary for each grade, however, if this is not furnished we feel that a method of sampling of the word offerings in each . , . e * ( ■ . 'V .'•> < , •. -1 . . r v ' : ; - ^ ' ’ ’i - f . . . . . . i./ ' •• - 138 - grade must be resorted to by the superintendent. The reliability of such a meth- od will depend upon the extent of the samplings made and the degree to which the words obtained in this manner are representative of all words in the grade vocab- ulary. Therefore, as suggested before, the textbook maker should protect his speller against chance methods of comparison by providing an alphabetical find- ing list of the word offerings in each grade. The sampling may be made by selecting the words, for example, in every f ourth page and checking them against the two dimensional check list selected for this purpose. (The Ayres -Teachers College list and the Anderson-Ashbaugh list are the only ones of any extent that are now available for purposes of devising such check lists.) To assist the superintendent in making comparisons of his findings for the spellers that he may examine, we are showing in Table XLII a conspectus of the rankings of the ten spellers examined by us. We are refraining from making any combinations because these rankings may mean different things for different people. For example, we might combine columns 1 and 2 which show, from two different points of view, the extent of the utilization of existing vocabulary studies. However, the two ideas are not the same, and the difference in ranks indicates this. Moreover, the value of r by the foot-rule formula amounts to only C,25» Again columns 4 and 5 might be combined because each represents the notion of repetition. Indeed the data represented by these two columns are so closely related that they tell practically the same story. Tftis will, likewise, be seen by inspecting the ranks entered in these two columns. The value of r by the foot-rule formula is approximat ely 0.85» A third combination could be made by using columns 6 and 7» Finally, it would be possible to combine the entire series into a general rank. In spite of these many possibilities for making combinations we did not make them, because we felt that the combinations would be made differently - - 1 < - .... . ... . : ■ . . ■ v' ; l 1 1 . ( • ■ • C t . . , . . . . . . 139 TABLE XL II. CONSPECTUS OF RANKINGS a ' OF SPELLERS Book Percent of Ayres- Teachers College words Percent of words in book that are Ayres- Teachers College words Difficulty of words Grade repe- tition of words Repetition of words for speller as a whole Check list Number of words in book Number of occur- rences 1 2 3 4 5 6 7 I 7.5 1 9 1 2 9 1 II 10 9 8 7 8 10 7 III 7.5 6 1 3 1 4 3 IV 3 5 5 5 6 6 5 V 6 10 7 4 5 5 6 VI 9 7 2 2 4 3 2 VII 4 4 3 8 7 7 8.5 VIII 1 3 6 6 3 2 4 IX 2 8 10 9 9 8 8.5 x * ✓ 2 r r~ t —i— i mtm i ■ r ~ 4 10 10 1 10 ( a ) ^Rank 1 is highest, rank 2, next highest, etc. - 14 C by different people and perhaps by the same person at different times. This ab- sence of agreement arises because of different purposes for which the combinations of rankings may be made. To assist the superintendent still further in his study and comparison of spellers, we are presenting in Table XLIII the different grade norms derived in our project. It would save the busy superintendent a great deal of time if he could secure already prepared a series of two dimensional check lists. Therefore, in order that we may render to him what service we can, we shall exhibit in the last pages of this chapter our series of check lists for Grades 2 to 8 inclusive so that he may use them if he cares to instead of taking the time to prepare a series, A statement or two concerning some of the words in our series of check lists may save the reader some confusion of thought. For example, he may wonder why we put the two words "girlie" and "auntie" in our lists. We must confess that these words were slightly shocking to our taste. However, after considering that they are real, dictionary words, that they are in the Anderson list and also in Ashbaugh's Iowa Spelling Scales, and that they were properly obtained by ran- dom sampling, we decided to take no liberties with them. Consequently, we left them in our check lists. * • < . . i ■ I , : < ( TABLE XLIII NORMS DERIVED PROM PRESENT PRACTICE 141 Median of eight spellers which offer dictation systematically* - 142 - 1 CHECK LIST FOR GRADE II. 1 Selected at random from the first 180 words of the Ander- son list that, according to the Iowa Spelling Scale, are spelled by not less than 27 nor more than 73 percent of second-grade pupils. about getting must than after girl name their again glad new them ago go ing night there also got no these am had note they because hear now thing been her only think before here or time best him order to better home our under came house place us car how please use card just price very city kind right want coming letter same we could long school well don't made send what down make should when each money so which ever month some who every more state wish fine mo st such work first Mr. sure would from much take write 1 - ^ - 143 - CHECK LIST FOR ®ADE III. Selected at random from the l8lst to the 540th inclusive words of the Anderson list that, according to the Iowa Spelling Seal are spelled by not less than 27 nor more than 73 percent of third- grade pupils. always enjoy amount enough another even answer fair aunt fall awful family balance farmer big folk busy following buy forward called found can’t gone che ck great children guess Christmas half church high county hoping cover however credit I'll date June dinner kindly do ctor large does later dollar mail enclo sing making mamma set market sick mind since Mrs. sold near something nearly stock November subject number summer October supply office therefore p ayment tho se people though per thought perhaps trusting pretty try quite until rate upon ready weather remember went reply whether report while request wrote seem yesterday sell young sending yourself . . . I' ■ ' . " •; . ■ :< c . .. 144 CHECK LIS? FOR GRADE IV. Selected at random from the 541st to the 900th inclusive words of the Anderson list that, according to the Iowa Spelling Scale, are spelled by not less than 27 nor more than 73 percent of fourth-grade pupils. advance daily item recently advertising dealer join regarding afr aid death knew replying aid decide loan require allow democrat local sale all right director lose secure answering discount madam September anyway district married serve awaiting education music shipped auto election national sooner beginning enjoyed ought stayed benefit entirely owner student between envelope package suit bought exchange paint superintendent buying excuse pair third careful fifteen placed thousand cashier finish policy tire cause finished prepared trying changed firm prize uncle color fully program unless complete greatly promptly vacation considerable heavy purpose waiting continue hello quality welcome contract important quit whole copies instead quote women . ■ , . ■: :w ' < ' 1 - . •A '' V ■ ' . • ' ... . . ■ l .U-': V.Uttl \ i ■ . ' ♦ • ■ • t ' 145 CHECK LIST FOR GRADE V. Selected at random from the 901st to the 1260th inclusive words of the Anderson list that, according to the Iowa Spelling Scale, are spelled by not less than 27 nor more than 73 percent of fifth-grade pupils. acquainted addressed advanced advice arrangement assist assure assuring attorney awfully basis buyer cedar circumstances clerk client clothe commercial common confidence connection convenient cooperation council crowd depot desired develop duplicate entitled enjoying error etc. fairly favorable February furnished garage gentleman hospital imagine in elude increase instant institution interesting invoice judgment knowledge lately lecture literature loss lovingly manner mentioned method missed model occasion offered operation opinion owing parcel parties patience planning practice prefer presence proceed product pro f easier, properly prospect purchase purchased putting refer registration regret remit requested running satisfaction satisfied securing shipping shown simply splendid surprise surprised system territory various vary view who se - 146 CHECK LIST FOR GRADE VI. Selected at random from the 126lst to the 1800 inclusive words of the Anderson list that, according to the Iowa Spelling Scale, are spelled by not less than 27 nor more than 73 percent of sixth-grade pupils.' ability acceptance accident actual actually advertise advertisement advised af f air allowed appeared appointed approval approved arranged arrival assigned assistance assumed assured bargain based bureau candidate cir cuit circulation combination concern conference considerably considered demo cratic design difficult difficulty disease distribution exact exactly exception expen sive familiar favored finally fortunate fraternally furniture grateful grippe independent indicate instance investigation invitation Latin liable limited manual maintain meant merely merchandise moral natural neglected neither operating patron practical preparation previous publication published readily reasonable referred regularly relative reliable renewal represented response responsible route sanitary secured semester senior science scientific similar so ciety spirit straight stopped strictly subscriber telephone urge usually CHECK LIST FOR GRADE VII. Selected at random from the 1801st to the 2340th inclusive words of the Anderson list that, according to the Iowa Spelling Scale, are spelled by not less than 27 nor more than 73 percent of seventh-grade pupils. acquaintance administration admission affectionately af fidavit algebra altogether alumni analysis announce announcement annually apparatus appearance applicant assessment assignment auditor ballot believing blizzard bonus brief capacity candidacy civics favorably quantities coarse generally quantity completely haste regretting compliment inconvenience safety consequently individual satisfy constitution inferior sense control influence signature cooperative institute sincere correspond instructor solicit correspondent jobber stationary coupon majority stomach crocheting necessity studying crowded opportunity succeed engineering orchestra succeeded enrollment organize suggested enthusiastic particularly supervisor essential permanent surplus established physical sympathy evidently planned tatting exceptional possession type exceptionally possibility unnecessary exclusively practically unusual executive principle variety explanation professional vicinity extremely qualities welfare . — , .. : - ■ ■ . . ■ ♦i ■ i V k w< *, ‘ . ■ v Lit Selected son list that, not less than accommodate accompanying accredited accuracy acquire administrator advisable agricultural allotment ambitious anniversary anticipate anticipating appendicitis appreciating associate auntie authority beneficial bicycle Canvass carnival characteristic commissioner confirmation - 148 - CHECK LIST FOR GRADE VIII. at random from the 2341st word to the end of the Ander- according to the Iowa Spelling Scale, are spelled by 27 nor more than 73 percent of eighth-grade pupils. conservatory consultation continuous conveniently cooperating cordial courteous customary definitely demonstration disappoint disappointment embarrassment employees endeavor enormous enthusiasm ere executed exhausted exhibition fundamental geometry girlie guarantee Hallowe’en hastily immense ingredients inquiries installment interfere legislation leisure librarian license losing materially rnatur ity mechanical mischievous mortgage nickel notary occasionally o ccurrence originally parallel peculiar peculiarities perceived pho sphorus physician pneumonia politician possess poultice preliminary privilege relieved remembrance resemblance responsibility Sabbath satisfactorily sufficiently surgery thesis tuberculosis unusually sensible soliciting specially specification straightened J . - . J / ... . ■I . ' >. . ■ I . - 149 - CHAPTER IX SUMMARY AND SUGGESTIONS A demand ia being made by some of our leading contemporary educators fer projects, similar to ours, in the analysis and measurement of the instru- ments of instruction, especially of the textbooks that are put into the hands of the pupils in our public schools. The number of words in the vocabularies of the ten spellers examined in our project ranges widely from 3968 to 59*5 and the average is 4900, From sociological and psychological considerations the vocabularies of eight of the ten spellers are too extensive. We found that the learning and reviewing load in general increases in the advancing grades. However, the word offerings vary widely among the ten spellers for a given grade, e, g, in Grade ^ one speller has 843 words and another i960, indicating absence of agreement among makers of spellers as to the proper learning and reviewing load that should be car- ried by the pupils in a given grade. We found, in general, that the Ayres list of so-called "thousand commonest words" is well represented in the ten spellers. The range in the number of appearances of these words is from 808 in one speller to 99^ in another. The words in the Teachers Cillege list of so-called "second and third thousand most frequently used" do not make such a fqyorable showing* Of the 2000 words in this list only from 1019 to 1323 appear in any one ef the ten spellers. On the basis of the Ayres -Teachers College list of 3000 words, we found that from one-half to two-thirds of the spelling time of the pupils is consumed in studying words that are not among those "most frequently used". By - 150 - an examination of the grade offerings of the Ayres -Teachers College words we dis- covered wide variation among the ten spellers. For example, in Grade 2 the range is 174 wsrde of this list and in Grade 7 the range is 613. The data indicate that learning and reviewing the Ayres -Teachers College words is cumulative to the end of the middle grades. It was noted that in advanced grades there is a de- ceasing percent of all the words in the grades that are common to the Ayres- Teachers College list, and consequently that the pupil as he advances from grade to grade is confronted with a decreasing proportion of the "mest frequently used words”. Here again we observed wide diversity of practice among the ten spellers. On the basis of the median percent for the grades the Ayres "thousand commonest words" were offered in much greater proportion than the Teachers College "second and third thousand most frequently used words." Using the best single index number of difficulty that we ars able to find it was discovered that the ten spellers as a whole are practically of the same difficulty. On the basis of the Ay res -Teachers College words common to each of the spellsrs it was found, as was to be expected, that in general the spelling difficulty of the words progresses from lower to higher grades. Inversions, how- over, were detected in some of the spellers. Further, the spellers vary widely in the difficulty of the words in a given grade. Generally speaking, the same is true when we considered either of the components of our synthetic liet. These are indications that no one speller may bee best suited to a group of grade pupils in a particular school system. In general the range of difficulty from grade median to grade median is small in comparison with the range in difficulty within a given grade as exhib- ited by the different spellers. That is, there is a narrow range among things purporting to be different and a wide range among things purporting to be the same. When the difficulty of the grade offerings of the spellers was measured 1 Ll 0 y^ laon J[. ith pupil abiU *y» it was seen that t he spellers are too difficult - 151 - ; in the lower gfades and too easy in the upper ones* The provisions for habit -formation by means of repetition of words show a diversity of practice. For the books as wholes one speller has 4.1 average occurrences per word while another has 1.4. That is, on the basis of repetition, other things oeing aqual,the first speller has made three times as much provision for the teaching of the words in its vocabulary than has the second one. On investigation of the provision for habit -formation within the grades, a wide range of practice was again found among the ten spellers, especially in the lower grades. However, when the upper grades were reached it was found that there is a level of low repetitions per word in all the spellers. That is, none of the speilersoraake such automatic provision for habit-formation in the upper grades, re- gardless of the fact that the laws of habit-formation function with upper-grade children as well as with those of the loweiv grade. The foregoing statements are true not only for all words but also for the Ayres-Teachers College words common to the grades. An investigation of the extent to which words appear in different grades of the spellers revealed, again, wide variation among the ten books in our list. This fact is significant because spelling tests show that a given word or words is misspelled in more than one grade. In the study of the correlation between the difficulty of words and the number of times they occur it was discovered that the easy words are quite as like-* ly to occur frequently as the hard words. Consequently, the spellers are tending to keep the "easy” words easy and the “difficult" words difficult. It is recog- nized that spelling difficulty may not be the same as learning difficulty. How- sver, at present because of lack of experimental data on the latter, we are assum- ing a high correlation between the two. Therefore, the classroom teacher must by various devices provide the pupils with many more experiences with the difficult words than the spellers automatically provide. By an investigation of the correlation between the repetition of the words in the spellers and their frequency of occurrence in adult writing, it was noted that eight of the ten epellere indicated relatively high correlation between these two uses of words. However, this fact does not necessarily mean that these spellers are best suited to the needs of the pupils, because by virtue of the fre- quency of occurrence of words in life their correct spelling tends to be taken care of in the undirected experience of the pupils. On the other hand, we need to teach many words because they are misspelled when they are used although this may be seldom. An examination of the forma of presentation of the words in the spellers revealed that the authors* stock in trade was of three kinds, namely, calumn form, column review form, and dictation form. On the one hand, these forms were used by all the authors and on the other they included the major portion of the words. That is, other forms such as word analysis and plural formation were scat- tsringly and sparingly used in the spellers. However, in the use of the three majs>r forme of presentation, as well as in the minor ones, we are compelled te state that the spellers exhibit wide variation in practice. Th© reader has no doubt noted - perhaps not without impatience - the fact that disagreement among textbook writers is constantly asserted. We have frequently used such expressions as ’’large variation”, ”wide range”, "diversity of practice”, and the like. The fact is that the writers of spellers agree in nothing so closely as in their profession of faith. They unite in saying that they exemplify the results of recent vocabulary studies. Yet they do so in such a variety of ways and with such differing degrees of success that in many impor- tant aspects diversity is the most typical single descriptive term to apply to the product of their labor. The question of whether agreement is desirable or not may be raised. With pupils of varying intelligence, environment, and heritage and in the hands of teachers of different abilities, textbooks undoubtedly must be different if they are to produce anything like the same results. But the point in respect to 1fc 88 ?„ 8 P 8ller8 * 8 that there is no apparent recognition of a particular type of I . 1 , \ a "-J . . ' ..... . * ... • ' • u ' J, I • i , . .. . ; ' >. . . : . . . ... : .. . • ii . • .v „ ... !. J.\. . t . ',t .... : • • • < ■ • • f . ' , , , • . 0 . ' J, . i (3 '. ■ . t i 'V . { , " . ■( ' t. , i - 'J ■ * * ■i \ ■ . . . , .. ' ■ : ' . H, '• - 153 - teacher or pupil for whom any of the books is especially appropriate. An author ie unwilling to say: "This book is recommended only for use by graduates of nor— mal schools or collages with pupils from good homes and of predominantly Anglo- Saxon stock." If an author wars willing to say such a thing, his publishers would prevent him. The result is that each speller is apparently intended for general use - i, s, for the purpose of producing, as nearly as may be, an automa- tic control of a useful vocabulary among all types of children and through the agen- cy of all types of teachers. It seems clear that this purpose cannot be realized even "on the aver- age" equally well by books that differ so essentially in content. If, for example the presentation of words in isolation ought to be reinforced by dictation exer- cises, then it is evident that the same measure of success will not be attained by two books, of which one has no dictation at all and the other 65 percent of such material. If frequent reviews are necessary for sufficient learning, it is equal- ly clear that the book which presents each word once cannot compete with a book which presents each word an average of four or five times. If an adjustment of difficulty of material to pupil ability is desirable, the book whose gradation most nearly conforms to what we know of the ability of pupils will be superior to one which either shoots over the heads of lower-grade children or falls far short of the ability of upper-grade children. We have no experimental evidence concerning the effects produced by different characteristics in books. When we have such evidence we shall become far more intelligent both in writing and criticizing textbooks. Meanwhile, it ie perhaps sufficient to point out the divergencies among existing texts and to ob- serve that with ana avowed common purpose they are probably as variable in the results they secure as they are in the measures they adopt. - 154 - Suggestions In the development of our project certain suggestions have come te mind that may be of practical value to the school people who are in actual contact with the public schools* We have also picked up a number of suggestions for fur- ther research projects that may be of interest to workers in the field of spel- ling. Suggestions to Superintendents It is not feasible for the superintendent to undertake such an extended project as the present one when he is passing judgment on the spelling texts that may be used in the school system. The time and energy involved make this prohibi- tive, especially for the superintendent outside the largest cities. However, he can use an abbreviated technic and certain of the materials that we used. The superintendent should equip himself with the best scientifically derived vocabulary obtainable at the present time, perhaps the Ayres -Teachers College list of three thousand "most frequently used words.'* Then he should equip himself with an alphabetical list of the word offerings in each grade in each speller under consideration. This is a difficult thing to do because spel- lers in general have not been provided by the authors with such a list. (Of the ten books that we examined only two made such a provision. ) Therefore, owing to the exceedingly difficult and practically prohibi- tive task of finding the complete grade offerings of spellers which is very es- sential in anything like an accurate comparison of spellers, we recommend strong- ly that no speller be considered for adoption unless it is provided with such a "finding list" for each grade. We hold unequivocally that it is an affront to the schools of the country to expect spellers or for that matter any other text- books to continue to be adopted without knowledge of their contents. To be sure, the makers of textbooks have not "wilfully and maliciously" kept as a hidden se- cret the contents of their books. As a matter of fact, school people have not demanded that a complete exhibit of the contents be made. Moreover, the analysis c i. . 1 v , l- ' i ' ■' v ■ » < ' 4 Kj . .•» O' - . ’ • . • ■ ■ ' 1 1 . . , • • . i . i . „• 1 . i .i 1 1 .i. ■ ’ . . . I , . ; . . x i. » , . . _ . . ...•*■*■ . V- ' J - . I , - L •. . • •- 1 , . .i ’ . -J • ^ 1 ( • v. 1 ' . » ;i j. ' ■ -■ . . <■: ■: " ■ • v . . i C •1. (> . ‘ •; ' 'if • '( • • •••'.' • '■ . .v ■ . c i. i - - J • > 0 • 1 , • c * - „ *,■ • - < . *■ .. V r I . . , • ■ ' ■ v‘ . ; ■ ... • . . . , ■■ . r , : 1 M a . 1 ." ' ’0 ' 1 l a "f J w ' . V, ( V ' . . . t,, i ' . ■ ; . v 'j-s ■ ‘ I. r ■■ - •«**« , . . . o 1 1 i ...... c ( - 156 - In preparing the list of 100 words they should be selected from the standpoint of difficulty and frequency of use. As a concrete illustration of a possible method of selecting and arranging such a series of two dimensional check lists, we wish to call the reader’s attention to Chapter VIII in which is de- scribed the methods of procedure, materials used, and the data obtained by us when w© developed a series of two dimensional check lists for comparing the word offerings in Grades 2 to 8 inclusive in these ten spellers. Further, to assist tho superintendent in his study of spellers we have assembled in the beforw men- tioned chapter a series of grade norms for spellers derived in the progress of our project. The superintendent should give one ©r more standard spelling tests (Ayres, Buckingham, or the Teachers College) and determine the spelling ability of the pupils in the different grades. These data may then be used to assist him in deciding which speller on the basis of the difficulty of the word offer- ings in the grades is best suited to the ability of the pupils in each grade. He may also determine the proportion of each speller devoted to the different forms of presentation. When the data on the selection, gradation, repetition, etc. are derived by the superintendent in the foregoing manner, he may then compare them with the norms an similar phases of spellers derived by us by much more elaborate and ex- tensive methods (See Chapter VIII). We are not implying that the foregoing pha- ses oi spelling suggested for examination by the superintendent are the only im- portant ones. However, we submit that they are among the most important and should be given a relatively weighty consideration in the comparison of spellers. Suggestions to Teachers feome suggestions that may be valuable for teachers have occurred to us during the development of this project. The teacher should assist the superin- tendent in every way possible in executing the foregoing comparison of spellers. He may help in checking speller words against selected comparison lists, and he _ j - 157 - may assist by writing words on cards and alphabetizing them. It is not altogether inconceivable that he may use his classes* provided they are in the middle and up- per grades* for part of this clerical work. One forcible suggestion came to us* namely, that the teacher must make con- scious provision for relatively much more repetition of the more difficult words in the vocabulary of the speller. This perhaps may be attacked best by first giving preliminary tests over the word lists for several lessons ahead and then apportioning the relative amount of repetition on the basis of the spelling per- formance of the pupils on these words. Owing to the present lack of data on the learning difficulty, the teacher will have to continue repetition and testing until the words have been learned so that the recall of the correct spelling i 8 efficient. The limitations in the variety of forms of presentation, also, require that the teacher use many more devices for presentation than are suggested by the spellers. Only in this way can he lend the spice of variety to the learning of the spelling of the words. Suggestions to_ Superintendents and Teachers on SP el ling ££« only Typical of what may be done in other subjects. The specific recommendations for spellers that we have made in the pre- ceding sections are only typical of the recommendations that should be applied tc other subjects such as arithmetic, histories, and geographies, if we but pos- sessed the requisite concrete data by means of which we might carry them out. For these latter subjects the superintendent and his teachers should set up con- crete data of their own pending the activities of research workers. In history, for example, they can decide what items ought to be found in history texts, and go through the texts to see if they are presented. In the comparison of history texts, therefore, they should look for facts , for example, "the three most important causes of the Civil War M , (These should be determined upon and written out beforehand), "the suppression of the limited states Bank - . , . , . ... . . . • . . . . 'll. ... J . • . ' . . - ' ; . c ■ ^ -I. v. • , r. • , . < < - . . ...... . . / . . . . a. r- . ... ;n , . j ' - , ... , . . c , . ' ■ . V , t C - ■ o - - -i ' v . ; - : " C ... i . . J , • • . , , • ' .. < Q , ' - ‘ ' • • • , - 158 - by Jackson," and "the Missouri Compromise." Such a procedure would give much more tangible data for purposes of comparison than would consideration of the texts under such captions as "Organization of subject matter", or "Meet the needs of children." Suggestions to. Publishers and Authors Spellers Many suggestions that may be valuable to publishers and authors of spel- lers came to us during the progress of our project. We shall take the liberty of giving some of the more important ones that have not escaped us. One of the serious difficulties that confronted us in our examination of the forms of presentation was that the authors did not use a common terminology in their headings and contexts. Each speller, as suggested many times before, should contain an alpha- betical finding vocabulary for each grade. With each word might well be shown the number of times it is repeated, its spelling difficulty, and whether new to the grade or not. Such data would aid in comparison of texts and be of inestim- able value to the teacher. More of the "second and third thousand most frequently used words" should be included in the spellers since we have fairly reliable data on these. The words selected for a given grade should be better suited to the known spelling ability of children of that grade. The repetition of the words in the spellers should have a positive relation to the difficulty of the words. In general the more difficult a word the more frequently it should occur. I Textbook publishers should employ an expert in analysis of texts and give him ample assistance in order that the contents and forms of presentation may be accurately and extensively determined. This would relieve the field agents of the task of making superficial analyses of the texts they are presenting to the school authorities for purposes of adoption. ■ . , ■ t t > . . . •. . ' . l^j) . ■ . - . I . . - . . ... $:?: i v c fl . - • [) - 159 - Suggestions to Research Workers in Spelling A few suggestions to research workers in spelling concerning some of the problems that arose in the course of the development of our project may not be too presumptuous* A comparison of the Ayree-Teacher s College list of three thousand with the Anderson-Ashbaugh three thousand words would doubtless yield many words not common to the two. Consequently a composite of the two would give an appreciably larger number of words whose spelling difficulty and frequency of occurrence in life have been derived. Therefore, many more of the words in the vocabularies of spellers could be studied than is possible at present. A study of the frequencies of the words common to the foregoing lists might raise the question as to the reliability of the data on this phase of the two vocabularies independently derived and occasion further investigation in this field* The difficulty and frequency of the Anderson-Ashbaugh three thousand words should be placed on the same basis as the corresponding items for Ayres- Teachera College list* This would expedite further research work with these two lists. Regardless of the monumental work of Jones, the question of the words used by children should be reopened* We need to know these words with their frequencies and their difficulties. Then these may be used conjointly with words derived from adult usage to construct the proper vocabulary to be taught to children. It was sharply forced upon us that we need to have a much more exten- sive list of words whose difficulty and frequency have been derived. This would have aided us immeasureably in our examination of the word offerings in the spel- lers. It would also aid the teachers and the makers of spellers* Careful and extended experiments should be carried out under school- room condition s to determine: (a) the number of new words and review words that . * ■ . '■ ■ . . * ‘ . - 16 © - I pupils of a given grade can master efficiently; (b) the optimum median difficulty of words that children of a given grade should have presented to them; (c) the number of repetitions, other things being equal, that are required in a given grade in order that the children may learn efficiently a given word (This should be extended to many words as rapidly as possible); and (d) the relative worth of different forms of presentation in order that we may pass better judgment on this phase of spellers. We need to obtain reliable data on the frequency of occurrence in life and the difficulty of abbreviations and proper names. Another valuable and interesting investigation would be to determine the effect of the recent scientific vocabulary studies on the selection and gra- dation of words in spellers. Our findings might be compared with the findings in similar analyses of spellers that have been published in certain decades prior to the beginning of these vocabulary studies. An interesting historical develop- ment in the selection and gradation of words might thus be worked out. In Conclusion In concluding the description of our project and the discussion of its findings we feel that many of our assumptions have defects which are obvious to the reader and that our technic of analysis and our modes of comparisons are crude. However, if we have contributed a modicum of data, suggested possible types of further analysis, developed slightly certain tools for research, given some impetus to the sorely needed analysis and measurement of the instruments of instruction, and reached reasonably rigorous conclusions concerning the phases of spellers examined, we shall not feel that our protracted labors have been wholly in vain. e| -rov , " ■ ■ . . ♦ * . 161 SELECTED BIBLIOGRAPHY Anderson, W. N. The determination of a spelling vocabulary based up on written correspondence . (Unpublished doctors' dissertation.) University of Iowa. 1917. Ashbaugh, E. J. " Iowa Spelling Scales ," University of Iowa Extension Bulletin, University of Iowa, 1919. Ayres, L. P. Measuring scale for ability in spelling . New York: Russell Sage Foundation, 1915* Ayres, L. P. "The application to tables of distribution of a shorter method for computing coefficients of correlation," Journal of Educational Research, April, 1920. — — ™ ~ " ‘ Ayres, L. P. "A shorter method for computing the coefficient of correlation," Journal of Educational Research , March, 1920. Buckingham, B. R. Spelling ability: its measurement and distribution . Teachers College, Columbia University, 1913". Buckingham, B. R. "Textbooks: their cost and improvement" (Editorial). Journal o f Edu cat io nal_ Re search, March, 1920. Buckingham, B. R. "Research in textbook publishing," (Editorial). Journal of E ducational Research , October, 1920. Cook, W. A* end O'Shea, M. V. The child and his spelling . Indianapolis: Bobbs- Merrill Company, 1914. Eldridge, R. C. Six thousand common English words . (Published privately.) Rochester, N. Y., 1911. Horn, Ernest. "Principles of method in teaching as derived from scientific investigation," Eighteenth Yearbook of the National Society for the Study of Education, Part II . Bloomington, Illinois: Public School Publishing Company. Judd, C. H. "Analyzing textbooks," Elementary School Journal , October, 1918. Jones, W. Franklin. Concret e investigation of the material of English spelling. University of South Dakota, Vermillion, South DakotaT 1915. ~ Knowles, Rev. J. The London point system of reading for the blind . 1904. Mead, C. D. "The best method of selection textbooks," Educational Administration and Supervision , February, 1918 . ~ - (Briggs and Kelley). "Sixteen spelling scales," Teachers College Record , September, 1920. Si ' " 162 - Starch, Daniel. Educational psychology . Now York: The Macmillan Company, 1919. Stoops, R. 0. "The use of score cards for judging textbooks," American School Board Journal , March, 1918 . — 1 — — ■ Thorndike, E. L. Mental and social measurement , second edition. Teachers Col- lege, Columbia University, 1913. Thorndike, E. L. "Recent developments in educational measurements, M Fifth Con- ference on Educati onal Measurements . (Bulletin of the Extension Division, Indiana University, 1918 .) Tidyman, W. F. The teaching of s pelling . Yonkers-on-Hudson: World Book Company, Woolfolk, Algar. The need of intensive work in spelling . (Unpublished). Woody, C. "Application of scientific method in evaluating subject matter of spellers," Journal of Educational Research . February, 1920. . , - • . . • . . . . - • • - 163 - VITA Arlie Glenn Capps, son of George A* and Ellen Capps, was born near Stahl, Missouri, March 2, 1 887 • His elementary education was received in the rural schools, and his high school and part of his undergraduate collegiate edu- cation was received in the State Teachers College, Kirksville, Missouri. After teaching in rural schools, he became in various places principal of elementary and secondary schools and city superintendent. In 1915 he entered the University of Missouri from which he received the B. S. degree in 1916 and the A. M. degree in 1917* The last two years that he attended the University of Missouri he held the Peabody Fellowship in Education, in 1917-18 he was director of the cooperative study of the rural schools of Missouri and statistician to the state department of education. Since 1918 he has been lecturer and graduate stu- dent in the College of Education of the University of Illinois. He taught at the University of Illinois during the summer session of 1919 and at the University of Missouri during the summer session of 1920. He has published ten articles among them being: "Cooperative study of the rural schools of Missouri,” 69 th Report of the Public Schools of the State of Missouri, pp. 36-95. "Curriculum content of a high school spelling course," Journal of Educational Research, Vol. 2, No. 3, pp. 626-35*