un:v^R3:Ty of calsfornea at los angeles r; LATIN IN SECONDARY SCHOOLS H. A. BROWN A Study of Ability in Latin in Secondary Schools A DESCRIPTION OF A METHOD OF MEASURING ABILITY IN LATIN, WITH A STATISTICAL STUDY OF THE RESULTS OF A SURVEY OF INSTRUCTION IN LATIN IN NEW HAMPSHIRE SECONDARY SCHOOLS By H. A. BROWN President State Normal School Oshkosh, Wisconsin PUBLISHED AT STATE NORMAL SCHOOL OSHKOSH, WISCONSIN 19 19 ^ I Copyrighted, 1920, by H. A. Brown PA, ^ PREFACE This investigation was begun when the writer was director of the Bureau of Educational Research connected with the New Hampshire Department of Public Instruction. A preliminary test was given in 1916. Most of the data contained in the present study were gathered near the end of the school year in 1917. The scoring of the papers and the tabulation of the data were done during the summer of that year. The study was made possible by funds with which to maintain the Bureau, granted by the General Education Board, for the pur- pose of carrying on scientific studies of educational practice. It should be stated that the General Education Board merely granted to the State Superintendent of Public Instruction the funds for maintaining the Bureau and in no way directed what investigations should be made. The Bureau acted entirely on its own initiative and accepts full responsibility for its findings. This monograph has been written for two classes of readers. There is a large class of people who are directly and very vitally interested in Latin as it is taught in both secondary school and col- ^ lege. They are mostly school and college teachers and adminis- trators, who desire to know the results of any investigations which throw light upon defects in present methods of teaching Latin, and they are also interested in any suggestions in the direction of im- proved methods of instruction. Their interest is thus chiefly in the practical side of the investigation. There is also a large class, made up of scientific investigators in education, who are interested not only in the practical results of the investigation, but also in the methods by which the results are secured. They will insist upon a full presentation of the data from which the conclusions were derived. For this reason the data have been presented in as great detail as the limits of the monograph will permit. This study was undertaken for purely administrative purposes to answer certain questions with reference to the success with which Latin was being taught in the secondary schools of the State as a whole. Therefore, in many cases, a general average gives a suffi- cient answer. Many tables and graphs which might have been III LO IV PREFACE printed arc omitted for this reason. It is believed by the writer that while this may make the monograph less interesting to the scientific investigator, it makes it more readable for the general school administrator. Enough data have been given so that any- one who desires to do so may check the accuracy of the work in every important particular. I desire to acknowledge here my indebtedness to several people who have assisted me in this study. To Professor Henry C. Morrison, of the School of Education, University of Chicago, formerly Superintendent of Public Instruc- tion for New Hampshire, I am indebted for valuable suggestions and criticism throughout the entire time during which the investiga- tion has been in progress. The study was made under his general direction. I am under especial obligation to Miss Margaret G. Kennedy, first assistant in the Bureau of Educational Research, for patient and untiring eflforts in giving the tests, for suggestions on the general plan of the study and especially for advice and painstaking labor in connection with the statistical work involved in evaluating the tests and in interpreting the results. Her assistance during the earlier stages of this study was invaluable. Mrs. Mabel A. Riordan, executive secretary. State Normal School, Oshkosh, Wisconsin, has rendered valuable service during the later stages of this study, especially in the final statement of conclusions. Professor Truman L. Kelley of Teachers College, Columbia University, has read the manuscript and oflfered many valuable suggestions. Dr. ?>. R. Buckingham of the University of Illinois has read a part of the proof. Dr. W. \V. Theisen and Dr. Carter Alexander of the Wisconsin State Department of Public Instruction have also read the proof. I am indebted to the General Education Board for its support of this investigation. State Normal School. H- ^- BROWN. Oshkosh, Wisconsin, July 1, 1019. TABLE OF CONTENTS CHAPTER PAGE I. Gp:neral Purpose and Nature of the Investiga- tion ... I II. Stecieic Purposes of the Investigation .... 3 III. The Tests and Their Application 6 IV. Development of the Tests 13 V. The Latin Sentence Tests 26 VL The Connected Latin Test 39 VII. Latin Grammar Test 57 VIII. Latin Vocabulary Test 67 IX. Time Devoted to Study of Latin 76 X. Method in Latin 78 XL Value of the Study of Grammar 85 XIL Relation Between Ability to Apprehend the Meaning of Latin and Knowledge of Con- struction 93 XIII. Ability in the Fundamentals of Latin in Rela- tion to Time Devoted to Latin Study .... 98 XIV. Evaluation of Method in Latin Instruction . 104 XV. One Obstacle to Success in Latin and the Remedy 117 XVI. Comparative Standing of Pupils in College . . 123 XVII. Character of the Pupils' English 126 XVIII. The Teaching of Latin 136 f. ■ Appendix A Distribution Tables for Latin Sentence Test B . 145 Appendix B Distribution Tables for Latin Sentence Test A . 149 Appendix C Distribution Tables for Connected Latin Test . .153 Appendix D Distribution Tables for Latin Grammar Test . . 162 Appendix E Distribution Tables for Latin Vocabulary Test . 166 Appendix F Conversion Table for P. E. Values 170 VI LIST OF TABLES TABLE PAGE 1. Distribution Showing the Number of Sentences Translated Cor- rectly in Each Year — Latin Sentence Test B 15 2. Number of Pupils in Each Year who Translated E^ch Sentence Correctly — Latin Sentence Test B 16 3. Per Cent of Pupils in Each Year who Translated Each Sentence Correctly — Latin Sentence Test B 17 4. Difference between Fifty Per Cent and the Per Cent in Each Year who Translated Each Sentence Correctly — Latin Sentence Test B 18 5. P. E. Equivalents of the Difference between Fifty Per Cent and the Per Cent in Each Year who Translated Each Sentence Correctly — Latin Sentence Test B 19 6. P. E. Intervals Shown between Consecutive Years in the Case of Each Sentence — Latin Sentence Test B 20 7. Distance of the Median of Each Year above Zero — Latin Sentence Test B 23 8. Location above Zero of Each Sentence — Latin Sentence Test B 24 9. Final Scale Values of Sentences — Latin Sentence Test B . . 25 10. Average Scores by Schools for Ability to Translate Latin Sen- tences — ^Latin Sentence Test B — ^Year I 26 11. Average Scores by Schools for Ability to Translate Latin Sen- tences — Latin Sentence Test B — Year II 28 12. Average Scores by Schools for Ability to Translate Latin Sen- tences — Latin Sentence Test B — Year III 28 13. Average Scores by Schools for Ability to Translate Latin Sen- tences — Latin Sentence Test B — ^Year IV 29 14. Class Averages for Record of Improvement in Terms of Scores Made by Pupils — Latin Sentence Test B 29 15. Distribution Table for Number of Sentences Translated Cor- rectly — Latin Sentence Test A 30 16. Number of Pupils in Each Year who Translated Each Sentence Correctly-^Latin Sentence Test A 32 17. Per Cent of Pupils in Each Year who Translated Each Sentence Correctly— Latin Sentence Test A 33 VII VIII LIST OF TABLES— Continued TABLE PAGE i8. P. E. Equivalents of the Difference between Fifty Per Cent and the Per Cent in Each Year who Translated Each Sentence Correctly — Latin Sentence Test A 34 19. Final Scale Values of Sentences — Latin Sentence Test A . . 35 20. Average Scores by Schools — Latin Sentence Test A — Year I . 36 21. Average Scores by Schools — Latin Sentence Test A — Year II . 36 22. Average Scores by Schools — Latin Sentence Test A — Year III . Z7 23. Average scores by Schools — Latin Sentence Test A — Year IV . 37 24. Record of Improvement for Latin Sentence Test A — Class Aver- ages in Terms of Scores Made by Pupils 38 25. Scale Values of Points of Key to Connected Latin Test ... 46 26. Distribution Table for Amount Attempted for Connected Latin Test 50 27. Distribution Table for Amount Correct for Connected Latin Test 51 28. Distribution Table for Comprehension for Connected Latin Test 52 29. Class Averages for Connected Latin Test — Year II ... . 53 30. Class Averages for Connected Latin Test — Year III ... . 54 31. Class Averages for Connected Latin Test — Year IV ... . 55 32. Class Averages for Record of Improvement in Connected Latin Test 56 Z2,- Distribution Table for Number of Constructions Correct — Latin Grammar Test 60 34. Number of Pupils in Each Year who Answered Each Construction Correctly — Latin Grammar Test 61 35. Per Cent of Pupils in Each Year who Answered Each Construc- tion Correctly — Latin Grammar Test 62 36. P. E. Equivalents of the Difference between Fifty Per Cent and the Per Cent of Pupils in Each Year who Answered Each Construction Correctly — Latin Grammar Test 63 Final Scale Values of Constructions — ^Latin Grammar Test . . 63 Average Scores by Schools — Latin Grammar Test — Year I . . 64 Average Scores by Schools — Latin Grammar Test — Year II . . 64 Average Scores by Schools — Latin Grammar Test — Year III . . 65 Average Scores by Schools — Latin Grammar Test — Year IV . . 65 38. 39 40, 41 42 Record of Improvement in Terms of the Scores Made by Pupils — Latin Grammar Test 66 LIST OF TABLES— Continued ix TABLE PAGE 43. Distribution Table for Number of Words Correct — Latin Vocabu- lary Test 67 44. Number of Pupils in Each Year who Gave the Correct Meaning of Each Word — Latin Vocabulary Test 69 45. Per Cent of Pupils in Each Year who Gave the Correct Meaning of Each Word — Latin Vocabulary Test 70 46. P. E. Equivalents of the Difference between Fifty Per Cent and the Per Cent of Pupils in Each Year who Gave the Correct Meaning of Each Word — Latin Vocabulary Test .... 72 47. Final Scale Values of Words — Latin Vocabulary Test ... y^ 48. Average Scores by Schools in Terms of the Scores Made by Pupils — 'Latin Vocabulary Test — ^Year I 74 49. Average Scores by Schools in Terms of the Scores Made by Pupils — ^Latin Vocabulary Test — Year II 74 50. Average Scores by Schools in Terms of the Scores Made by Pupils — Latin Vocabulary Test — Year III 74 51. Average Scores by SchooL«^ *i Terms of the Scores Made by Pupils — Latin Vocabulary Test — Year IV 75 52. Record of Improvement in Terms of Scores Made by Pupils — Latin Vocabulary Test 75 53. Time Devoted to Study of Latin 77 54. Comparison of Efficiency in Knowledge of Construction — Year I 86 55. Comparison of Efficiency in Knowledge of Construction — Year II 87 56. Comparison of Efficiency in Knowledge of Construction — Year III 88 57. Comparison of Efficiency in Knowledge of Construction — Year IV 89 58. Comparison of Efficiency in Knowledge of Construction — Summary 90 59. Relation between Ability to Apprehend the Meaning of Latin and Knowledge of Construction — Year II 94 60. Relation between Ability to Apprehend the Meaning of Latin and Knowledge of Construction — Year III 95 61. Relation between Ability to Apprehend the Meaning of Latin and Knowledge of Construction — Year IV 96 62. Relation between Ability to Apprehend the Meaning of Latin and Knowledge of Construction — Summary 97 63. Time Allotments and Efficiency for Pupils who Have Studied Latin One Year 99 64. Time Allotments and Efficiency for Pupils who Have Studied Latin Two Years 100 X LIST OF TABLES — Continued TABLE PAGE 65. Time Allotments and Efficiency for Pupils who Have Studied Latin Three Years loi 66. Time Allotments and Efficiency for Pupils who Have Studied Latin Four Years 102 67. Time Allotments and Efficiency — Summary 103 68. Comparison of Different Methods for Ability in the Fundamentals of Latin— Year I 105 69. Comparison of Different Methods for Ability in the Fundamentals of Latin — Year 11 106 70. Comparison of Different Methods for Ability in the Fundamentals of Latin— Year HI 107 71. Comparison of Different Methods for Ability in the Fundamentals of Latin— Year IV 108 72. Comparison of Different Methods for Ability in the Fundamentals of Latin — Summary 109 "i- Comparison of Different Methods in Relation to Time Allot- ments — ^Year I no 74. Comparison of Different Methods in Relation to Time Allot- rri/ents — Year H in 75. Comparison of Different Methods in Relation to Time Allot- ments — Year HI 112 76. Comparison of Different Methods in Relation to Time Allot- ments — Year IV 113 77. Comparison of Different Methods in Relation to Time Allot- ments — Summary 114 78. Present Overlapping of Classes and Way in which Pupils Ought to be Classified According to Ability, Based on Connected Latin Test 118 79. Present Overlapping of Classes and Way in which Pupils Ought to be Classified According to Ability, Based on Latin Sentence Test A 120 80. Standing of Pupils in College 124 CHAPTER I GENERAL PURPOSE AND NATURE OF THE INVESTIGATION One of the very important needs in secondary education at the present time is that current practices and methods should be evalu- ated in a scientific manner. Secondary educational practice in the past has been to a degree a blind, undefined and aimless procedure, and these characteristics still remain in a large measure. Even at the present time the validity of the purposes which control and guide instruction and the choice of subject-matter, in some respects, is doubtful. We still have altogether too vague notions concerning the real reasons why we teach at all such studies as algebra, literature and Latin and not something else in their places. If it were neces- sary to prove definitely just what specific values in terms of a valid social economy many subjects now in the program contribute, under our present forms of teaching, it would be very difficult to justify their existence or to defend the main objectives which those who teach them have in mind as the chief results to be secured. Plans for reorganizing the work of the secondary school on a broad scale, both as to content and method, are under discussion.^ The present type of secondary education is being subjected to vigor- ous criticism in many quarters on account of its lack of definiteness of aim, the alleged superficiality of much of its work and the non- functioning relationship to life of many things which pupils are required to study. It is on account of this marked tendency at pres- ent to examine anew, critically, the whole process of instruction, aims, organization and program-material that it is highly desirable to find out exactly where we stand with regard to efficiency in the 1 For a detailed discussion and criticism of modern secondary education, see the following, which are good illustrations of the modern critical attitude. Snedden, David. New Problems in Secondary Education, School Review, Vol. XXIV, No. 3, pp. 177-187. Morrison, H. C. Reconstructed Mathematics in the High School, Thir- teenth Yearbook of the National Society for the Study of Education, Part I, pp. 9-31. Flexner, Abraham. A Modern School, Publications of The General Edu- cation Board, Occasional Paper No. 3. 2 LATIN IN SECONDARY SCHOOLS various subjects. In a word, we need to have very definite knowl- edge of just the degree to which and the points at which we are succeeding and failing according to our present plans and methods. Only when we have this information can we make a beginning most intelligently in devising new methods of instruction, discarding old subjects and introducing new ones, reorganizing program-content and establishing new aims. It is especially imperative in this scien- tific age that we should not proceed blindly to throw out a method of instruction which has been in use for a long time merely because we think it is not producing results and replace it by another which we think may be more eflfective in this direction. There is an especially great need at the present time of this evaluating process in all of the older school subjects. Do we know in Latin whether the present results are such that there is a need for devising new methods of instruction? Are we sure as to just those particulars in which changes in method are desirable? The answer to these questions is that we have practically no objective evidence but many poorly grounded opinions and beliefs. Preliminary Nature of the Investigation The investigation described in this monograph was entered upon for the purpose of securing some light upon these problems. It pre- tends to be only a preliminary attempt in a comparatively new field. The method used is crude in some respects from a scientific and statistical point of view and is not intended to be in any sense a final and best way of measuring ability in Latin. The results are admitted to be tentative and provisional. The necessity of getting some reasonably definite information concerning actual results of teaching in this field is so great that it seems better to proceed at once with somewhat imperfect methods, than to delay until an entirely satisfactory technique of measurement has been developed. It is believed that the most rapid progress can be made in working out final and valid methods of measurement by gaining as wide experience as possible with such crude plans as can be devised at the present time provided, of course, that they give reliable results. It is felt that the method herein described is accurate enough to be very serviceable and to furnish information of great value. CHAPTER II SPECIFIC PURPOSES OF THE INVESTIGATION The general purpose of this investigation, as previously stated, was to test and evaluate present aims, methods and practices in connection with the teaching of Latin in secondary schools. In addi- tion to this general aim, a number of detailed and specific problems presented theinselves for solution. Before proceeding to a descrip- tion of the test and a discussion of the results, a statement of these particular problems should be made. Among the most important the following may be mentioned. (i) It is very desirable to know exactly to what extent the schools are nozv succeeding in teaching a mastery of the Latin lan- guage. Do pupils acquire a competent reading knowledge of Latin at the end of four years of instruction? Is there any objective evi- dence that high school boys and girls acquire a genuine appreciation of the literature of the language in the original or that they are capable of doing so? Do they master the common aspects of Latin grammar? Are the schools securing real results along these lines or only a smattering? (2) What is the best method by which to teach Latin? There are in the State a good many different kinds of procedure. In two schools the direct method was in use for several years in the first- year work. In others the so-called grammatical method is the basis of instruction, with variations in emphasis upon different aspects of the work in various localities. Several schools are using a method to be described later which has been called the transla- tion method. How do the results by these different methods com- pare? If some are in a marked degree superior to others, that fact needs to be made known. What are the chief factors in method which produce efficiency? Is method of any particular importance or does ability on the part of pupils depend upon something entirely different? If so, upon what? (3) What is the relation of knowledge of grammar and vocab- ulary to ability to translate Latin? What correlations are there between these abilities? What do they signify? Are the pupils who 4 LATIN IN SECONDARY SCHOOLS have been thoroughly drilled in Latin grammar in a formal way superior in speed and accuracy in getting the thought of Latin ? (4) The method by n'hich knozvlcdge of construction may be acquired best is important. In some schools a great deal of time is spent on formal study of grammar, discussion of facts of linguistic construction and syntactical analysis of the texts studied. In other schools very little of this is done. Do the classes in which this large amount of time is given to grammatical study show any greater superiority in knowledge of construction than those in which much less or indeed very little time is devoted to it? (5) Ability in Latin in relation to the time devoted to the sub- ject has been too long neglected. Some schools require nearly twice as much time to be spent by pupils in the study of Latin as do others. How do the efficiencies of these different classes compare? Is there any waste of effort here? \\'hat time allotment in each year is sufficient to secure satisfactory results? (6) Units of measurement, standard tests and norms of ability are needed in Latin. These have been devised in considerable num- bers for certain subjects but up to the present time only a begin- ning has been made in Latin. This is a subject in which it ought not to be difficult, eventually, to develop adequate means of measure- ment which may be applied by school administrators as a part of modern scientific school supervision. In view of this fact, the slow progress which has been made along this line is more to be deplored. An important purpose of the present investigation is to secure data with which to make a beginning in the establishment of norms of ability in Latin. It is impossible to judge our present effi- ciency in Latin except in terms of some such definite, objective standards. We need to find out what results are being obtained and what should be attained. We shall have to know what results are possible of attainment under normal conditions. With this knowledge and only with it, can we decide what a satisfactory product is in any school. (7) Are present purposes and aims in the teaching of Latin valid? We shall first need to find out what aims now control the teaching. The question then will be. Are they valid aims? If not, what should be the main objectives of four or five years of time PURPOSES OF THE INVESTIGATION 5 spent in the study of Latin by pupils, the teaching effort devoted to the subject and the money expended by the community? A com- plete study of this aspect of the problem will not be undertaken. Certain obvious aspects of it only will be discussed. All of these and many other similar matters are exceedingly important questions and they await the light which experimental investigation can throw upon them. It has been possible up to the present time to answer them only in terms of opinion which has small scientific value. It has not been possible in this study to answer all of these questions. Limitations of Purpose of This Study It should be understood distinctly by the reader of this monograph that the purpose of this study was to investigate only certain aspects of the work in Latin. It should be borne clearly in mind that it was no part of the aim to determine the educational value of Latin, its place in the curriculum or whether it should be in the course of study at all, the best time to begin the study of Latin or problems of that sort. The purpose of this study was primarily to determine the efficiency of the work in Latin in the secondary schools of a single State under present methods of instruction. CHAPTER III THE TESTS AND THEIR APPLICATION General Plan of the Tests In connection with any test the question which arises at the out- set is, What abihties need to be measured in order to make known the complete amount, degree, or extent of ability possessed by an individual or a group of individuals? We have our problem then in the form of the question, What specific abilities in Latin must be measured in order to furnish adequate information about a pupil's total ability? A second question of considerable importance is, To what extent are these abilities measurable? In reading English it is easy to see that important things to know, which are easily measurable, are the amount which can be read in a given unit of time and the degree of comprehension of the reader. Any plan which will give reasonably accurate information along these two lines will prove of great practical value in establishing norms of ability, evaluating methods and diagnosing class and individual needs. In Latin it seems evident that the speed and correctness with which an individual is able to get thought from a passage of the language is one important measure of his ability. In this investigation it is assumed to be true that the person who can get the largest amount of thought, and get it most correctly, from a given selection of Latin in a unit of time, is the most efficient regardless of excellencies or deficiencies in other minor particulars. Is not this the end for which Latin is taught? Other aspects of Latin to which a good deal of time is given in teaching are construction and vocabulary, but they are of secondary importance as aspects for measurement. They are sufficiently important, however, to deserve attention. Probably these three abilities are the most funda- mental. Several difi^crent tests were used in measuring the pupils' ability. They are described below. Connected Latin Test The test which was used as a connected Latin test in this inves- tigation consisted of a rather easy passage of Latin. It presented THE TESTS AND THEIR APPLICATION 7 no particular difficulties of vocabulary, construction or thought, which ought to be at all troublesome to pupils who have studied the subject two years. The pupils were given a certain specified time in which to get and write as much of the thought of the Latin as possible. They were scored on the amount which they could write in English in the given time. This test was given to 1,160 pupils, of whom 582 had studied Latin two years, 317 three years and 261 four years. What the Test Measured The test was designed to measure the pupils' ability to react to the total situation presented by a connected passage of Latin. It determined their ability to interpret the meaning of the Latin and express it in writing. It was thought that this kind of a test might be better than a scale in the form of a series of isolated sentences of increasing difficulty. According to the writer's understanding of the term, to read means carrying the thread of a story or argument to a climax or conclusion, understanding its meaning and grasping the significance of the whole. It involves interpreting a passage of discourse in terms of its thought content. In the absence of evi- dence there was the possibility that the ability required to react in this way to a connected passage might be something quite different from that required to give the thought of a series of isolated sen- tences which, when thus detached from their settings, are more or less meaningless. In any case, it was thought desirable to try both plans. Therefore it was decided to devise also a Latin test consist- ing of a series of isolated sentences of increasing difficulty. The latter is somewhat easier to score and if it is as adequate a test of ability to interpret the meaning of Latin discourse, that fact should be known. It was held that it might be entirely possible that the ability to read as defined above may be measured indirectly by a Latin sentence test in an entirely satisfactory manner just as tem- perature is measured indirectly by the height of a column of mer- cury which varies as the temperature varies. This was one of the important problems connected with this study. No preconceived notions were entertained but tests were devised to find the facts. The results are discussed later in this bulletin. 8 LATIN IN SECONDARY SCHOOLS Latin Sentence Tests The Latin Sentence Tests consisted of a series of Latin sentences, the first of which was very easy and the others of which were increasingly more difficult. The first few sentences were so easy that it was expected that pupils who had studied Latin a few weeks would be likely to succeed with them. The most difficult sentences were thought to be such that only exceptional pupils who had had Latin in secondary schools at least four years could deal with them successfully. A plan was adopted for evaluating the sentences so as to have them range by about equal steps of difficulty from the easiest to the most difficult. It was found when the test was first given in the schools that it did not fulfill all of the expectations. It was discovered at once that there were not enough easy sentences. The test as originally arranged consisted of twenty sentences. When the defect noted above was discovered, ten easier sentences were added. The test as composed of the twenty sentences originally chosen was called Test B and that consisting of the final thirty sentences was called Test A. Test A was given to 813 pupils divided as follows: first year, 371 ; second year, 233; third year, 104; fourth year, 105. Test B was taken by 2,160 pupils, of whom 942 were in the first year, 598 in the second, 347 in the third and 273 in the fourth. Both tests are given on a following page. The tests were so printed that the pupils could write the translation of each sen- tence directly under it. Time enough was allowed so that each pupil could do all that he was able to translate before he had to cease work. The time allowed on Test A was forty minutes and on Test B thirty. It was found by actual experience that this time was suflficient to allow a pupil to do all that he was able to translate. The Latin Grammar Test A Latin Grammar Test was given to 1.974 pupils, of whom 715 had studied Latin one year, 591 two years, 364 three years and 304 four years. This, in a word, measured the pupils' ability to name and describe Latin constructions. The test appears on a later page. It is believed that to be able to name and describe Latin con- structions in a formal sense may be a very different thing from abil- THE TESTS AND THEIR APPLICATION g ity to react correctly to constructions in translation. It happens many times that a pupil can translate correctly a given construction whenever he meets it in its functional relationship in a sentence and yet be very ignorant with respect to such formal knowledge con- cerning it as would be necessary in naming and describing it. For that reason it is very desirable that a functional grammar test should be devised. This functional grammar test should measure the success of the pupils in reacting to certain grammatical con- structions in the material which they have to translate. It is assimied by the writer, in the absence of evidence to the contrary, that this formal knowledge of grammar may bear little relation to functional knowledge. Although attempts were made, a satisfac- tory test of functional knowledge of grammar was not devised. The Latin Vocabulary Test The Latin Vocabulary Test was given to 841 pupils, of whom 509 had had one year of Latin, 211 two years, 117 three years and 104 four years. All that has been said concerning formal and functional gram- mar tests applies with equal force to vocabulary. It is believed that there may be a great difference between the ability to give one cor- rect meaning of a word, when a list of the words is given, and the ability to react correctly to a word in a sentence, i. e., to give a meaning which is proper in the particular context. No satisfactory test for functional knowledge of vocabulary was constructed, although several attempts were made. Information Secured From Schools Detailed information, as called for on the form below, was col- lected from each school : LATIN REPORT Dear Certain information concerning the work in Latin in. 10 LATIN IN SECONDARY SCHOOLS is asked for below. It will be a service to this Bureau, which will be greatly appreciated, if you or your Latin teacher will furnish the information and return this folder at once. A stamped envelope is enclosed. Town or City Date of Test School Latin Teacher Numerals below indicate years of Latin study: I, first year pupils; II, second year pupils, and so on. II III IV 1. Total number of minutes per week of class work in Latin, on the part of pupils, in each class (Put answers under right numerals to indicate the proper year in school.) 2. Number of minutes per week devoted to study of Latin, by pupils in each class, outside of class period, in school or out 3. Pages of translation of connected Latin during the present year by pu- pils in each class 4. Average size of Latin classes in each year (Divisions taught by one teacher.) 5. To what extent is sight translation a part of your method? (This question may be answered by stating the proportion of all the translation which is at sight.) 6. Which of the methods named below corresponds most closely to that which you use? (Answer as indicated below.) I use method A, B, C (cross out two) in the first year. I use method I, II, III (cross out two) in the years above the first. METHOD IN FIRST YEAR: A. Direct method. No systematic study of grammar during the year. Grammatical principles learned incidentally through use of the language in conversation or in translation or reading. No study of beginners' book con- taining grammar lessons. Translation, reading and conversation are learned from the beginning in a functional way by practice on material adapted to the stage of progress of the pupils. B. Translation method. No systematic study of grammar from a book. Daily translation of sentences from blackboard. Grammatical constructions THE TESTS AND THEIR APPLICATION ii taught concretely in sentences. Grammar learned incidentally through exten- sive translation. Easy translation from books begun soon after the beginning of the year and continued throughout the year. No regular and systematic study of a beginners' book. C. Grammatical method. Study and completion during the year of a beginners' book which contains a systematic presentation of the principles of grammar together with Latin sentences to translate and English sentences to write in Latin. This book is the basis of the year's work. METHOD IN YEARS ABOVE THE FIRST : I. No systematic study of grammar except as principles are explained to the extent needed to understand the meaning of the language, — when met in translation, — or learned incidentally by reacting to them frequently in translation. No systematic study of grammar in connection with prose com- position, to the extent of assigning lessons to be studied in a text on grammar. II. Grammar studied in connection with prose composition once a week or the equivalent. Grammar references assigned to be studied and recited in connection with the prose lessons. No direct attention given to grammar except in connection with prose composition. III. Systematic study of grammar both in connection with prose com- position and the Latin texts studied in class. Regular grammar lessons assigned for study and recitation. Constant attention to principles of syntax in the texts studied, by questions on the constructions met and otherwise. 7. Has this same method been in use during the period that the highest class now studying Latin has been in school? If not, describe the method by which the classes tested have been taught in each year. Please give a rather complete statement. 8. If the method used does not correspond closely to any of the above, give on the space below a detailed description of your method. 9. Is there any study of grammar after the first year? If so, by what method and to what extent? (Please make a rather complete statement here.) 10. Do you mtake use of perception-card drill on words or groups of words? If so, where in the course and to what extent? 11. Do you require the writing of English sentences in Latin during the first year ? To what extent ? 12. To what extent, and how, is prose composition taught above the first year? 13. Do you give regular tests in Latin ? If so, please descirbe their nature and state the extent to which they have been given. 14. Describe any other significant aspects of your method which are not covered by the above questions. Namie of person making this report : Address — ^Town or city : 'Street : State : 12 LATIN IN SECONDARY SCHOOLS Each pupil was asked to put the following information on the back of each test paper : To be filled out by pupils : Name Date City or town State School Class Age : years months This is my year of Latin study. Schools in which I have studied Latin each year: I II Ill IV Names of Latin teachers in each year: I II Ill IV Every school has been visited one or more times by the writer or by some other member of the State Department of Public Instruc- tion, and the methods of teaching in use have been studied critically and other information has been secured through correspondence.^ 1 Previous to taking up the directorship of the Bureau of Educational Research, the writer served as Deputy Superintendent of Public Instruction for the State of New Hampshire, and devoted his attention chiefly to second- ary school inspection, in which a survey was made of the instructional aspects of the work of these schools and a large amount of significant material was accumulated in the form of several large volumes of typewritten reports. For example three weeks were spent studying the work of one high school, a week making a detailed survey of another, and so on. The work in Latin in nearly all of the schools involved in this investigation, therefore, had already been studied thoroughly before the test was given. Two other Deputy Superintendents, for four years, w-ere engaged in the same sort of inspection and reporting, in connection with the work of the secondary schools, and the writer had free access to their reports and numerous conferences with them. The State Superintendent of Public Instruction had been in very close touch with the secondary schools for more than a decade and has contributed invaluable information concerning the history of the work in Latin in each school. For these reasons, the Bureau has been able to have at its disposal unusually complete and accurate knowledge concerning the instruction in Latin in the State. CHAPTER IV DEVELOPMENT OF THE TESTS In the following discussion only a very brief treatment of the theory underlying the statistical method applied in the derivation of the scale values of these tests has been given. Anyone who is not particularly interested in the theory of the matter may pass over this chapter. Theory of Scale Derivation^ The discussion of the method of developing the tests and the theory which underlies it is based on Latin Sentence Test B This was given to the greatest number of pupils. This is hardly large enough on which to base the evaluation of a scale, but on the other hand, it is believed that the final scale values secured are very much more accurate than any mere arbitrary approximations which could be adopted. The method used in developing and evaluating the tests is the now familiar method Gsed by Thorndike, Buckingham, Henmon, Trabue, Woody, and others. To any one who is familiar with their methods our indebtedness to them will be apparent on every page of this chapter. Since the theory on which the evaluation of the tests 1 For a complete discussion of the theory referred to in this chapter see the following: Woody, Clifford. Measurements of Some Achievements in Arithmetic. Bureau of Publications, Teachers College, Columbia University, New York City. Trabue, IM. R. Completion Test Language Scales. Bureau of Publications, Teachers College, Columbia University. Buckingham, B. R. Spelling Ability : Its Measurement and Distribution. Bureau of Publications, Teachers College, Columbia University. Thorndike, E. L. Introduction to the Theory of Mental and Social Meas- urements. Bureau of Publications, Teachers College, Columbia University. Henmon, V. A. C. The Measurement of Ability in Latin. Journal of Educational Psychology, Vol. VIII, pp. 51S-538 and pp. 589-599- Van Wagenen, M. J. Historical Information and Judgment in Pupils in the Elementary School. Bureau of Publications, Teachers College, Columbia University. 14 LATIN IN SECONDARY SCHOOLS is based is discussed so thoroughly in their writings and elsewhere, only so much of it wall be given here as is necessary for an under- standing of its application to the problem of developing these Latin tests and scales. \\' ith so many almost identical developments of the same theory in print, one would be hardly justified in giving an extensive description of the method. Therefore, this chapter is con- fessedly sketchy. The method which has been used in evaluating these tests is based upon three assumptions. The first of these is that that is diffi- cult upon which a large number of pupils fail and easy upon which a large number succeed. It is assumed, also, that ability in Latin is distributed according to the normal probability curve. The third assumption is that the variability in Latin ability is the same in each of the four high school years. For any further discussion of these three assumptions the reader is referred to the studies mentioned in the footnote on the previous page. The unit of measure used in connection with these tests was the probable error or median deviation, as it is commonly called. The probable error has been defined as follows: "It is an amount of ability above or below the median such as to include half the total number of ratings. In a normal distribution, if we arrange all of the participants in the order of their ratings, and if we count from the beginning one quarter of the way and three quarters of the way, half the difiference between the measures which we reach is the probable error." In a normal distribution surface, if a perpendicular is so drawn that on each side of it are found fifty per cent of the cases, the point at which this perpendicular intersects the base line of the surface of frequency is the median point. If perpendiculars are erected on each side of the median at such distances that between the median and each of the perpendiculars just twenty-five per cent of the sur- face is included between the median ordinate and each of the other two ordinates, the distance along the base line from the median to either ordinate is the probable error distance. This gives us a very convenient measure for scaling our Latin sentences or words or con- structions. Of course it is understood that the base line of the normal dis- tribution surface and the curve are asymptotic and theoretically DEVELOPMENT OF THE TESTS 15 never meet. It is customary in studies of this kind to assume that they meet at some convenient distance from the median. In this study it is assumed that they meet at 4.6 P. E. TABLE 1 ABILITY TO TRANSLATE LATIN SENTENCES Latin Sentence Test B Distribution Table for Number of Sentences Translated Correctly Number of Sentences Year I 50 1 449 2 278 3 86 4 48 5 15 6 5 7 7 8 2 9 ID I 11 12 I 13 14 15 16 17 18 19 20 Number of Pupils Tested 942 Median Number of Sentences Translated Correctly 1-938 25 Percentile i-4i3 75 Percentile 2.746 Quartiie 667 Number of Pupils Year II Year III Year rV 13 2 • • • 60 15 9 132 32 23 121 49 25 98 45 31 55 46 38 47 47 33 2i 44 23 15 21 23 6 18 19 9 9 19 7 9 3 5 3 8 I 3 9 3 I 4 2 I 5 I 2 I 598 347 273 2,-777 5.663 6.318 2.580 3-770 4-363 5-445 7.551 8.989 1-433 1.440 2-313 From the above it is seen that any Latin sentence may be located at a point on the base line of the surface of frequency. The point at which it is located will indicate its degree of difficulty. All meas- ures of the difficulty of sentences will be expressed in terms of P. i6 LATIN IN SECONDARY SCHOOLS E. All distances are measured to the right or left from the median. Distances to the left are negative and those to the right are posi- tive. A sentence which was translated correctly by ninety-five per cent of the pupils would be located on the easy end of the scale and near the end. On the other hand, a sentence translated correctly TABLE 2 ABILITY TO TRANSLATE LATIN SENTENCES Latin Sentence Test B Number of Pupils in Each Year Who Translated Each Sentence Correctly Sentence Tear I Year II Year III Year IV 1 539 Z'i^2 208 148 2 237 305 217 156 3 58 191 162 141 4 40 149 1 19 104 5 48 147 127 89 6 27 117 loi 65 7 I 49 92 58 8 I 28 65 57 9 o 19 42 50 10 3 25 35 37 11 I 24 33 30 12 II II 24 31 13 2 4 4 88 14 I 8 23 35 15 10 21 30 16 I 17 21 20 17 5 14 8 18 o 3 14 8 19 2 II 20 o o I 3 by only five per cent of the pupils would be on the difficult end of the scale. In a word, each sentence would be located at such a point on the base line that the percentage of the surface at the right of an ordinate erected at the point in question would be equal to the per- centage of the pupils who translated the sentence correctly. Thus a sentence translated by seventy-five per cent of the pupils would be located at — i P. E. A sentence translated correctly by twenty- five per cent of the pupils would be at -f-i P- E. These distances DEVELOPMENT OF THE TESTS 17 are expressed as distances from the median. The sentence which is located at — i P. E. would be at that distance below the median and the one located at -j-i P. E. would be at that distance above the median. In finding the scale value of a sentence it is necessary to know the per cent of deviation from the median of the per cent TABLE 3 ABILITY TO TRANSLATE LATIN SENTENCES Latin Sentence Test B Per Cent of Pupils in Each Year Who Translated Each Sentence Correctly Sentence Tear I Tear II Tear III Tear rv 1 57.2 52.2 59.9 54-2 2 25.2 51.0 62.6 57-1 3 6.2 31.9 46.7 51-6 4 4-2 24.9 34.3 38.1 5 5.1 24.6 36.6 32.6 6 2.9 19.6 29.1 23.8 7 0.1 8.0 26.5 21.2 8 0.1 4-7 1^7 20.9 9 ... 3.2 12.1 18.3 ID 0.3 4-2 lo.i 136 11 0.1 4.0 9-5 1 10 12 1.2 1.8 6.9 1 1.4 13 0.2 0.7 1.2 25.4 14 0.1 1.3 6.6 12.8 15 1.7 6.1 ii.o 16 0.1 2.8 6.1 7-3 17 0.8 4-0 2.9 18 0.5 4-0 2.9 19 ... 0.6 40 20 ... 0.3 I.I of pupils who translated the sentence correctly. In Table i is given a distribution of the number of sentences translated correctly in each year, and in Table 2 the number of pupils translating each of the sentences is given. Table 3 gives the per cent of pupils in each year who translated each of these sentences correctly, Table 4 the dif- ference between fifty per cent and the per cent of pupils in each year who translated each sentence correctly. Table 5 the P. E. equivalents of the difference between fifty per cent and the per cent in each year i8 LATIN IN SECONDARY SCHOOLS who translated each sentence correctly. These P. E. equivalents have been obtained directly from a conversion table. These P. E. values indicate the location of each sentence on the base line of the normal probability surface above or below the median. It will be noticed that the sentences do not have the same location, with ref- TABLE 4 ABILITY TO TRANSLATE LATIN SENTENCES Latin Sentence Test B Difference Between Fifty Per Cent and the Per Cent of Pupils in Each Year Who Translated Each Sentence Correctly Sentence Year I 1 7-2 2 —24.8 3 -43.8 4 -45.8 5 —44.9 6 —47.1 7 —49.9 8 —49.9 9 10 —49.7 11 —49.9 12 —48.8 13 —49-8 14 —49.9 15 16 —49.9 17 18 19 20 erence to the median, for each year, but have different scale values for the different years. The next problem, then, is to develop a general scale with a scale value for each sentence which will deviate as little as possible from the values for each of the years. In order to do this, two things must be known: (i) the distance between the consecutive year medians; (2) and the location of a common zero point. Year II Year III Year IV 2.2 9-9 4.2 l.O 12.6 7-1 —18. 1 — 3-3 1.6 —25.1 — 15.1 —1 1.9 —25.4 —13.4 —17.4 —30.4 — 20.9 — 26.2 — 42.0 — 23.S —28.8 —45.3 —31.3 — 29.1 —46.8 —37-9 —3^-7 -45.8 —39.9 —36.4 — 46.0 —40.5 —390 —48.2 —43.1 -38.6 —49.3 —48.8 — 24.6 —48.7 —43.4 —37-2 -48.3 —43.9 —39.0 —47-2 -43.9 —42.7 —49.2 — 46.0 —47.1 —49.5 — 46.0 —47.1 —49.4 —46.0 . . . —49.7 —48.9 DEVELOPMENT OF THE TESTS 19 In measuring the distance between the year medians a number of different values are sometimes secured. The averages of these, with certain of them weighted, have been taken usually as the inter- vals between the various years. (i) One measure of the year interval is usually found by deter- mining the difference in the position of each sentence in relation to TABLE 5 ABILITY TO TRANSLATE LATIN SENTENCES Latin Sentence Test B P. E. Equivalents of the Difference Between Fifty Per Cent and the Per Cent in Each Year Who Translated Each Sentence Correctly Sentence Year I 1 — 0.269 2 0.991 3 2.281 4 2.562 5 2.425 6 2.811 7 4.600 8 4.600 9 10 4.083 11 4.600 12 3346 13 4.275 14 4.600 15 16 4.600 17 18 19 20 "ear II Year III Year IV 0.082 —0.372 —0.156 0.037 —0.476 —0.265 0.698 0.123 —0.059 1. 005 0.600 0.449 1.019 0.508 0.669 1.269 0.816 1. 057 2.083 0.931 I.186 2.483 I.318 1. 20 1 2.746 1-735 1.340 2.562 1.892 1.629 2.597 1.944 1,819 3.III 2.199 1.788 3,643 3.346 0.982 3300 2.234 1.685 3.146 2.293 I.819 2.834 2.293 2.155 3-571 2.597 2.81 1 3.820 2.597 2.81 1 .... 3-275 2.597 • • ■ • 4.083 3-395 the medians of two consecutive years. By way of illustration, if we refer to Table 5, we find that the first sentence is situated .269 P. E. below the median of Year I and .082 P. E. below the median of Year II. We have here a difference of .187 P. E. Table 6 gives these differences for the various sentences. The average of all the dif- ferences for the interval between Year I and Year II gives one 20 LATIN IN SECONDARY SCHOOLS measure of the interval between these years. This is called the composite average. (2) Another measure of the interval between the years has often been obtained by the so-called quartile method. "If we have a normal surface of distribution the quartile of any dis- tribution should be equal to the P. E. of that distribution. There- TABLE 6 ABILITY TO TRANSLATE LATIN SENTENCES Latin Sentence Test B P. E. Intervals Shown Between Consecutive Years in the Case of Each Sentence Sentence Interval I-II 1 —0.187 2 1.028 3 1-583 4 1-557 5 I-406 6 1.542 7 2.517 8 2.117 9 10 1. 521 11 2.003 12 0.235 13 0.632 14 I-300 15 16 1.766 17 18 19 20 Interval II-III Interval III-IV 0.290 — 0.216 0.439 — 0.21 1 0.575 0.182 0.405 0.151 0.511 0.161 0.453 0.241 1. 152 0.255 1. 165 0.1 17 I.OII 0.395 0.670 0.263 0.653 0.125 0.912 0.41 1 0.297 2.364 0.166 0.549 0.853 0.474 0.541 0.138 0.974 0.214 1.223 0.214 .... 1. 128 • . ■ ■ 0.688 fore, if we divide the quartile of a distribution into the crude score intervals, we will get the interval between the medians of the grades [in this study the years] in terms of P. E. Since for each interval between the grades [years] there are two quartile meas- ures, the average of the two quartiles is used as a divisor of the crude score interval between the grades [years]." DEVELOPMENT OF THE TESTS 21 (3) A third method, commonly called the distribution method, is in common use for determining the intervals between the me- dians of the different years. This is based on the overlapping of the year distributions. There are pupils in every year who equal or exceed the medians of several of the next higher years and it is true, also, that some pupils in a given year will fail to reach the median achievement of the years below. None of these three methods was found well adapted to this test. The measure of the distance between the consecutive year medians which has been used in this study is that obtained by the so-called sentence method. The measures of the intervals obtained from the select group only have been used. That is, as a measure of the in- tervals between the years, use has been made of the average of those determinations only which come from values between — 2 P. E. and -)-3 P. E. in Table 5. The reason for this may need explanation. The easy sentences should not have as great weight as the more difficult in determining the intervals between the years. Since the distribu- tions for Years I and II have long tails, the percentages lying be- tween the medians of Years II, III, and IV do not differ by amounts as large as would be the case with normal distributions. This re- sults in values which are too small in the case of the lower indirect and the upper indirect determinations of the intervals between years obtained from overlapping of distributions. This skewness may be attributed to the difficulties of the particular sentences given and not to a real skewness in ability, so that the values in question are genuinely too small to represent the true situation. As a further reason, the uneven difficulty of certain of the sentences for the dif- ferent years would seem to make it desirable to use the intervals ob- tained by the sentence method with values from the select group only. The sentence method of determining intervals assumes a normal distribution for the talent of pupils, but it does not assume that the scores on a particular test follow a normal distribution. The quartile method assumes this latter fact, as well as the former, and the overlapping method is not entirely free from the particular combination of difficulties found in this test. For these reasons the sentence method with values obtained from the select group only is held to be the soundest method. 22 LATIN IN SECONDARY SCHOOLS The measures of the intervals between the consecutive year medians obtained by the sentence method with only select group values taken are as follows : Interval Interval Interval I-II II-III III-IV I-I55 .655 .204 One more step must be taken before we can determine our scale values. We must locate a zero point in order to be able to know just how much more difficult one sentence is than another. The method of determining the zero point is as follows. It was found by the data which are set forth in Table i that 44.71 per cent of the pupils of Year I are between those who did not succeed in translating one sentence correctly and the median for that year. By transmuting this per cent into terms of P. E. we have 2.397 P- E. This means that the median of Year I is 2.397 above zero. In the same way we find that 47.8 per cent of the pupils in Year II are between those who did not translate any sen- tences correctly and the median of Year II. This means that the median of Year II is 2.986 P. E. above zero. We already know that the median of Year II is 1.155 P. E. above that of Year I. If now we subtract the distance that the median of Year II is above Year I from the distance that the median of Year II is above zero, we find how far the median of Year I is above zero. This gives us 1.831 P. E. The zero referred to here is of course the arbitrary zero. It should be explained that the arbitrary zero is very different from the absolute zero, which means just not any ability at all in the thing in question. A pupil who fails to get any of the sentences of this test right does not necessarily have zero ability in Latin in the absolute sense. The arbitrary zero point as used here means the inability to translate correctly a single sentence under the particular conditions connected with the test.^ Now it will be remembered that in a normal distribution the P. E. and the quartile are equal. If we divide the median of Year I by the quartile we shall have another measure of the distance that the median of Year I is above zero. This is found to be 2.906 P. E. By dividing the median of Year II by the quartile of the same year 1 For a complete discussion of this idea see Thorndike's Introduction to Mental and Social Measurements, p. 16 f f. DEVELOPMENT OF THE TESTS 23 we get a measure of the distance that the median of Year II is above zero. Then, if we subtract the distance that the median of Year II is above the median of Year I from the distance that the median of Year II is above zero we get the distance of the median of Year I above zero. This is found to be 1.481 P. E. If we take the average of the above four determinations we shall have a satisfactory measure of the distance of the arbitrary zero point below the median of Year I. The four determinations are as follows, with their average : From the distribution of Year 1 2.397 From the distribution of Year II 1-831 From the achievement of Year 1 2.906 From the achievement of Year II 1.481 Average 2. 1 54 Thus, from the above, we know that the distance of the arbi- trary zero point is 2.154 P. E. below the median of Year I. Since we have the determination of the distance of the median of Year I above the arbitrary zero point, it is an easy matter to find how far above zero the medians of each of the other years are. This is given in Table 7. TABLE 7 ABILITY TO TRANSLATE LATIN SENTENCES Latin Sentence Test B Distance of the Median of Each Year Above Zero. Year Above Zero Below Next Tear I 2.154 I-IS5 n 3.309 0.655 III 3.964 0.204 IV 4.168 We already know the P. E. values of each sentence for each year from Table 5. We need now to refer all of these to the zero point in order to get a general value for each. In finding the loca- tion above zero of each sentence it is necessary to add to or sub- tract from the values given in Table 5 appropriate values from Table 7. To illustrate, the first sentence in Table 5, Year II, has 24 LATIN IN SECONDARY SCHOOLS a value of — .082. Table 7 shows that the median of Year II is 3.309 P. E. above zero. If now we subtract .082 from 3.309 we shall have Z-227y which is the distance above zero of the first sentence. By this procedure all the values in the column for Year II in Table 8 have been obtained. By a similar method those for the other col- TABLE 8 ABILITY TO TRANSLATE LATIN SENTENCES Latin Sentence Test B Location Above Zero of Each Sentence Sentence "^ Year I Y^ar II Year III Year IV 1 1.885 2>-227 3-592 4012 2 3145 3-272 3488 3-903 3 4-435 4-007 4087 4-109 4 4-716 4-314 4564 4-710 5 4-579 4-328 4.472 4-837 6 4-965 4-578 4.780 5-225 7 6.754 5-392 4895 5.354 8 6.754 5-792 5.282 5-369 g 6.055 5-699 5-508 10 6.237 5-871 5-856 5-797 11 6.754 5-906 5-908 5987 12 5.500 6.420 6.163 5-956 13 6.429 6.952 7-310 5-150 14 6.754 6.609 6.198 5-853 15 6.455 6.257 5987 16 6.754 6.143 6.257 6.323 17 6.880 6.561 6.979 18 7-129 6.561 6.979 19 7.689 6.765 20 8.047 7.563 umns have been determined. If the values in Table 5 are positive, we add instead of subtracting. From the data of Table 8 it is now possible to determine the final scale values of our sentences. In previous studies by this statistical method it has been customary to weight certain of the values in Table 8. The reason for doing this is that it has been felt "that those values which come from those distributions where the median achievements were farthest from the location of the problem [sentence] should have little or no weight." Following DEVELOPMENT OF THE TESTS 25 this practice, values in Table 8 have been given double weight when the sentence is less than i P. E. from the median achievement of that year and single weight if it is more than i P. E. but not more than 4 P. E. If it is more than 4 P. E. distance from the median achievement of that distribution, it is disregarded. The TABLE 9 ABILITY TO TRANSLATE LATIN SENTENCES Latin Sentence Test B Final Scale Values Sentence Scale Value 1 3-i8 2 345 3 412 4 427 5 459 6 4.87 7 5-13 8 5-49 9 5-75 10 5.84 Sentence Scale Value 11 5-93 12 ". . . . 6.01 13 6.14 i 14 6.22 15 6.23 16 6.24 17 6.81 18 6.89 19 7-22 20 756 average of those that are not disregarded is taken as the final scale values of the sentence. These are given in Table 9. For the other tests only the scale values are given, together with enough of the basic tables to enable anyone who desires to check the accuracy of the work to reconstruct all of the tables. CHAPTER V THE LATIN SENTENCE TESTS Nature of the Tests The Latin sentence tests consisted of Test A and Test B as pre- viously noted. These are both made up of a series of sentences of known difficulty, ranging from easy to difficult. The use of Test B is not advised for pupils who have studied Latin but one year, for TABLE 10 ABILITY TO TRANSLATE LATIN SENTENCES Latin Sentence Test B Average Scores by Schools YEAR I School I . 26 . 6 . 19 . 7 • 27 • 3 • 12 . 16 . 4 • 9 • 18 . 24 . 17 . 31 • Average Score 2.03 ..•• 2.35 2.76 .... 3.02 .... 2.87 .... 3.06 •••• 323 .... 3.00 .... 3.76 4.61 .... 3.20 •••• 332 .... 324 .... 3-41 .... 3.44 School 2 . 14 . 21 . 22 . ID . 33 ■ 13 • 23 . ir . 32 ■ 29 . 34 • 5 . 8 . 20 . Average Score 3.42 3.25 3-05 3.38 3-94 3-50 3-51 3-51 341 363 3.63 3-70 3S3 4.09 4.1 1 the reason that it is clearly too difficult for first year pupils. When only upper class pupils are to be tested, it will be found serviceable. Method of Scoring In scoring these sentence tests the pupil's mark is the scale value of the hardest sentence translated correctly. Thus if a pupil trans- lated correctly all of the first ten sentences, his score is the scale LATIN SENTENCE TESTS 27 value of the tenth sentence. The pupil is, however, given credit for any sentences that he may do beyond the point at which he ceased to translate continuously. For example, if the pupil translated cor- rectly all of the first ten sentences and then translated correctly the thirteenth and fourteenth, his score would be the scale value of the twelfth sentence. There are sound arguments for claiming that this is a fair method of scoring. LATIN SENTENCE TEST B 1. Puella est parva. 2. Est copia frumenti in agris nostris. 3. Hoc facto, pueri discesserunt. 4. Eis militibus fuerunt scuta gladiique. 5. Dixit aedificium in quo Hannibal esset multos exitus habere. 6. Postero die collem fossa trium milium passuum munivimus. 7. Helvetii existimabant eam civitatem quae nuper pacata esset non bono animo esse. 8. Ante noctem statuendum est quid faciendum sit. 9. Cum Helvetii bello clarissimi essent, Caesar iter per provinciam dare recusavit. 10. Hoc iter tanto difficilius est ut melius sit alteram temptare.. ir. Postridie pugnandi causa itinere converso nostros sequi coe- perunt. 12. Multi Servium imperio prohibere cupiunt. 13. Quinque cohortes castris praesidio relinquit. 14. Priusquam pugnemus, arti militari studeamus. 15. Caesar profecturus Romam non exspectavit. 16. Caesar senatus in eum beneficia commemoravit, quod rex appel- latus esset. 17. Hannibal magnas copias comparat, quibus Italiam in potesta- tem suam redigat. 18. Primo quaeremus quae consilia probata sint. 19. Ut aegro, dum anima est, spes esse dicitur, sic ego, quoad Pompeius in Italia fuit, sperare non destiti. 20. Ita fit ut adsint propterea quod ofiicium sequuntur, taceant autem idcirco quia periculum vitant. 28 LATIN IN SECONDARY SCHOOLS TABLE 11 ABILITY TO TRANSLATE LATLM SEMTEXCES Latin Sentence Test B Average Scores by Schools YEAR II School Average Score 3 2.46 6 3-35 4 3.38 9 3-57 26 319 19 347 7 3-52 23 382 18 3.83 31 3-93 17 3-99 25 4-10 10 6.05 24 3-95 13 3-99 32 390 12 4.10 School Average Score 2 3-99 14 384 30 415 21 4-25 I 538 16 4.25 S 442 34 4.23 8 340 22 4.69 II 4-52 33 463 20 4.43 27 484 15 5-29 29 541 TABLE 13 ABILITY TO TRANSLATE LATIN SENTENCES Latin Sentence Test B Average Scores by Schools YEAR III School 3 • 6 . 2 . II . 9 • 34 • 7 • 4 ■ 18 . 30 . ID . 8 . 21 . 15 • 23 • 26 . Average Score ..-. 351 .... 448 .... 3.61 .... 4-52 .... 4.04 4.20 .... 4.27 •••• 4-35 ■••■ 4-35 .... 4.38 . ... 4.49 . ... 4.41 .... 4.46 . ... 4.27 . ... 4-52 .... 4.58 School 32 . 12 . 14 . 20 . 19 . 31 . 25 . 17 . 22 . 24 . 16 . I . 5 . 13 . 27 ■ 29 . Average Score .... 4.65 .... 4.71 .... 4.75 .... 4.83 .... 4.75 .... 4.79 .... 4.77 .... 484 .... 4.81 .... 4.90 .... 5.01 5.22 .... 5-14 . ... 5-40 .... 5.75 6.04' LATIN SENTENCE TESTS 29 TABLE 13 ABILITY TO TRANSLATE LATIN SENTENCES Latin Sentence Test B Average Scores by Schools YEAR IV School 21 . 26 . 3 • 19 . 18 . 14 . 4 • 8 . 12 . 6 . 32 . 24 . 20 . 17 • 16 . 27 . 22 . Average Score .... 3.89 .... 4.15 .... 4-30 .... 4.41 .... 4-31 .... 4.28 .... 4.46 .... 4.42 .... 4-56 .... 4.43 .... 4.71 .... 4.66 .... 5.60 .... 4-86 .... 4.85 .... 4.87 .... 4.98 School 23 ■ 2 . 9 • 31 . 10 . 25 . I . 30 . 15 . 7 . 34 . 11 . 13 . 33 ■ 5 . 29 . Average Score .... 5.02 .... 4.95 .... 4.99 .... 5.19 .... 5-24 .... 5.14 .... 5.34 .... 5.21 .... 5.39 .... 5-47 .... 5-13 .... 5-88 .... S.77 .... 5.84 .... 5-8o .... 5-93 TABLE 14 LATIN SENTENCE TEST B Record of Improvement In Terms of the Scores Made by Pupils CLASS AVERAGESi Year I . II . Class Averages .... 3.29 .... 4.09 Year III . IV . Class Averages 4.62 .... 4.86 1 These class averages are the averages of all the individual marks of pupils of the year in question. For example, in Year I the class average of 3.29 means that all of the individual scores of all pupils in this year when averaged gave this figure as the result. The individual score of a pupil is not the number of sentences which he translated, but it is the score which he obtained by the method described on the first page of Chapter V. 30 LATIN IN SECONDARY SCHOOLS TABLE 15 ABILITY TO TRANSLATE LATIN SENTENCES Latin Sentence Test A Distribution Table for Number of Sentences Translated Correctly Number of Sentences o I 2 3 4 5 6 7 8 9 10 II 12 13 15 i6 17 i8 19 20 21 22 23 24 25 26 27 28 29 30 Number of Pupils Year I 9 28 54 43 40 31 26 22 25 21 16 13 22 8 6 4 I I Number of Pupils Tested Median Number of Sentences Correct 25 Percentile 75 Percentile Quartile 371 Yearll 2 2 8 9 14 15 15 14 14 23 20 21 22 15 9 7 8 2 I 3 3 4 2 233 Year III Year IV I 2 5 6 7 6 8 10 5 7 9 II 8 6 5 I I 3 2 104 3 I 7 6 6 6 9 7 13 3 6 9 3 6 4 4 6 2 I 105 5-371 10.025 12.143 14.346 3-407 6.550 8.667 11.042 9.012 12.807 15.000 17.972 2.803 3-129 3-166 3-465 LATIN SENTENCE TESTS 31 LATIN SENTENCE TEST A 1. Puella est parva. 2. Via per oppidum est longa. 3. Puella cantat. 4. Est copia frumenti in agris nostris. 5. Homines pugnare parant. 6. Quis hie est? 7. Nauta stellam videt. 8. Earn legionem ad montem duxit. 9. Nocte castra movebo. 10. Has feminas laudo, quae bene laborant. 11. Miles amico sagittam dat. 12. Hoc facto, pueri discesserunt. 13. Dixit aedificium in quo Hannibal esset multos exitus habere. 14. Puer bonus a matre non saepe culpabitur. 15. Eis militibus fuerunt scuta gladiique. 16. Postero die collem fossa trium milium passuum munivimus. 17. Helvetii existimabant eam civitatem quae nuper pacata esset non bono animo esse. 18. Postridie pugnandi causa itinere converso nostros sequi coe- perunt. 19. Cum Helvetii bello clarissimi essent, Caesar iter per provinciam dare recusavit. 20. Multi Servium imperio prohibere cupiunt. 21. Ante noctem statuendum est quid faciendum sit. 22. Hoc iter tanto difficilius est ut melius sit alterum temptare. 22,. Priusquam pugnemus, arti militari studeamus. 24. Caesar profecturus Romam non exspectavit. 25. Caesar senatus in eum beneficia commemoravit, quod rex appel- latus esset. 26. Quinque cohortes castris praesidio relinquit. 27. Hannibal magnas copias comparat, quibus Italiam in potestatem suam redigat. 28. Primo quaeremus quae consilia probata sint. 29. Ut aegro, dum anima est, spes esse dicitur, sic ego, quoad Pompeius in Italia fuit, sperare non destiti. 30. Ita fit ut adsint propterea quod officium sequuntur, taceant autem idcirco quia periculum vitant. 32 LATIN IN SECONDARY SCHOOLS TABLE 16 ABILITY TO TRANSLATE LATIN SENTENCES Latin Sentence Test A Number of Pupils in Each Year Who Translated Each Sentence Correctly Sentence Year I Year II 1 325 202 2 199 210 3 200 17s 4 147 174 5 154 146 6 193 124 7 168 138 8 131 149 9 137 152 10 97 no 11 134 lOI 12 28 94 13 36 14 14 90 78 15 23 59 16 20 25 17 6 8 18 o 20 19 8 14 20 3 18 21 3 72 22 5 26 23 I 13 24 7 8 25 o 9 2^ 3 4 27 I 19 28 o 64 29 O O 30 O I Year III Year IV 94 91 100 98 90 100 91 89 72, 86 70 92 69 88 76 80 79 87 75 91 53 71 64 72 51 47 28 41 38 49 27 36 29 36 15 24 18 21 15 27 24 32 19 28 9 27 12 21 8 12 2 63 6 13 5 5 2 II 4 7 LATIN SENTENCE TESTS 33 TABLE 17 ABILITY TO TRANSLATE LATIN SENTENCES Latin Sentence Test A Per Cent of Pupils in Each Year Who Translated Each Sentence Correctly Sentence Year I Year II Year III Year IV 1 87.6 86.7 88.7 86.7 2 53-6 90.1 94-3 93-3 3 53-9 751 84.9 95-2 4 39.6 747 85.8 84.8 5 41-5 62.7 68.9 81.9 6 52.0 53.2 66.0 87.6 7 45-3 59-2 65.1 83.8 8 35-3 63.9 71.7 76.2 9 36.9 65.2 74.5 82.9 ID 26.1 47.2 70.8 86.7 11 36.1 43-3 50.0 67.6 12 7.5 40.3 60.4 68.6 13 97 27.4 48.1 44.8 14 24.2 33.5 26.4 39.0 15 8.9 25.3 35.8 467 16 5-4 30.9 25.S 34.3 17 1-6 10.7 27.3 34.3 18 0.0 1 1.2 14.2 22.9 19 2.2 6.0 17.0 20.0 20 0.8 7.7 14.1 25.7 21 0.8 6.0 22.6 30.5 22 1.3 3.4 17.9 26.7 23 0.3 3.9 8.5 25.7 24 1-9 3-4 II-3 20.0 25 0.0 8.6 7.5 11.4 26 0.8 1.7 1.9 60.0 27 0.3 S.6 57 12.4 28 0.0 3.0 47 4.8 29 0.0 0.0 I.I lo.s 30 0.0 0.4 3.8 6.7 34 LATIN IX SECONDARY SCHOOLS TABLE 18 ■ ABILITY TO TRANSLATE LATIN SENTENCES Latin Sentence Test A P. E. Equivalents of the Difference Between Fifty Per Cent and the Per Cent in Each Year Who Translated Each Sentence Correctly Sentence Tear I Year II Year III Year IV 1 — 1713 —1.649 —1.795 —1.649 2 — 0.134 — 1909 — 2.344 — 2.222 3 —0.145 —1.005 —1-531 —2.468 4 0.391 —0.986 —1.589 —1.524 5 0.318 —0.480 —0.731 — 1.351 6 — 0.074 — 0.119 — 0.612 — 1. 713 7 0.175 —0.345 — 0-575 —1.462 8 0.559 —0.527 —0.851 —1.057 9 0.496 —0.579 —0.977 —1.409 10 0.949 0.104 — 0.812 — 1.649 11 0.527 0.250 — 0.677 12 2.314 0.364 —0.391 —0.719 13 1.926 0.891 0.071 0.194 14 1.038 0.632 0.936 0.414 15 1-997 0.986 0.539 0.123 16 2.384 0.782 0.977 0.600 17 3.182 1.843 0.895 0.600 18 1.803 I-589 i-ioi 19 2.986 2.305 1.415 1.248 20 3.571 2.114 1.595 0.968 21 3-5"i 2.305 1. 115 0.756 22 3.300 2.706 1 .363 0.922 23 4.083 2.614 2.035 0.968 24 3-077 2.706 1.795 I-248 25 2.026 2.134 1-788 26 3.571 3.146 3.077 —0.376 27 4083 2.357 2.344 I-713 28 2.789 2.483 2.468 29 .... 3-077 1-859 30 3938 2.631 2.222 LATIN SENTENCE TESTS 35 TABIiE 19 ABILITY TO TRANSLATE LATIN SENTENCES Latin Sentence Test A Final Scale Values Sentence 1 . . . 2 . . . 3 ••• 4 ... 5 ••• 6 ... 7 ••• 8 ... 9 ... 10 .. . 11 . . . 12 . . . 13 ... 14 ... 15 ... Scale Value . . . 2.00 . . . 2.12 . .. 2.43 ... 2.79 ... 3-II . .. 3-12 ... 3-17 . .. 3.20 ... 3-34 ... 342 . .. 3.60 ... 3-94 ••• 4-45 • •• 4-54 . . . 4.60 Sentence Scale Value 16 17 18 19 20 21 22 2i 24 25 26 27 28 29 30 4.86 523 5-54 5.56 5-6o 5-70 5-71 5-80 5-91 6.03 6.05 6.18 6.22 6.77 6.97 Conclusion Concerning Ability to Translate Latin Sentences When we consider the abihty of the pupils involved in this inves- tigation to translate simple Latin sentences, certain facts concerning the work in Latin in these secondary schools stand out with great clearness. Tables 10-14 set forth the results secured from the use of Latin Sentence Test B. The sentences of this test, with one or two exceptions, are of the grade of those found in Latin begin- ners' books in use at the present time. When the test was made up, it was supposed that it would prove too easy for third and fourth year pupils in secondary schools. Table i shows the results for the entire group of pupils who took the test : In the first year the median number of sentences translated correctly was 1.93, in the second year, 3.77, in the third year, 5.66, and in the fourth year, 6.31. It must be remembered that on this test sufficient time was allowed for practically all pupils to translate all that they could translate anyway. This certainly is an inadequate result to secure from four years of instruction in 36 LATIN IN SECONDARY SCHOOLS Latin : a median ability to translate about six of the easiest of twenty sentences of the degree of difficulty of the easier of those found in an ordinary beginners' book. TABLE 20 ABILITY TO TRANSLATE LATIN SENTENCES Latin Sentence Test A Average Scores by Schools YEAR I School I . 12 . 26 . 7 • 6 . i8 . 4 • 21 . Average Score 2.05 .... 2.08 .... 2.38 .... 2.54 .... 2.58 .... 2.93 .... 2.86 .... 2.55 School 24 . 2 . 14 . 13 ■ 17 ■ 10 . 5 • Average Score .... 3.10 .... 3.II .... 3.20 ...■ 3-45 •••• 3-45 .... 3.48 .... 3.90 A glance at Tables 14 and 24 serves to impress the fact of the small amount of progress in Latin under current methods during the four years that it is studied. It will be understood that the class averages in Tables 14 and 24 for each year are the averages of the individual scores of all pupils and not the average number of sen- tences translated correctly. The average performance of all the pupils in Year I on Test B is 3.29. This is slightly more than TABLE 21 ABILITY TO TRANSLATE LATIN SENTENCES Latin Sentence Test A Average Scores by Schools YEAR II School 26 . 6 . 4 • 24 . 14 . 7 . ID . Average Score 2.63 . . . . 2.90 3.00 ... 3.19 .... 3.22 .••• 3.92 • •. 3-31 School 13 • 18 . 17 . 21 . 1 . 2 . 5 • Average Score .... 3.47 ■••• 3-57 ■■.. 3-65 .... 3-74 .... 3.56 .... 2,-77 ■■■■ 4-35 LATIN SENTENCE TESTS 37 TABLE 32 ABILITY TO TRANSLATE LATIN SENTENCES Latin Sentence Test A Average Scores by Schools YEAR III School 13 • 6 . 2 . 7 • 21 . 18 . 14 . Average Score .... 5.08 .... 341 ■••• 3-73 ••.. 3-75 .... 3.83 .... 3.90 .... 3.91 School 10 . 26 . I . 4 . 24 . 5 . Average Score .... 3.90 .... 3.96 ••.. 3-97 .... 4-25 .... 444 .... 4.90 the scale value of the easiest sentence, which indicates that a very considerable number of pupils failed to get even one sentence cor- rect. In the second year the class average is slightly less than the scale value of the third sentence. There seems to be only very slight improvement from the end of the second year to the end of the third year, for the average of all the pupils in the third year is about equal to the scale value of the fifth sentence. From the end of the third year to the end of the fourth year there is little improvement. In a word, the average for all pupils in the fourth year is not greater than the scale value of the sixth sentence. It must be remembered that we are speaking here of the sen- tences at the easier end of the scale. The entire four years of Latin TABLE 23 ABILITY TO TRANSLATE LATIN SENTENCES Latin Sentence Test A Average Scores by Schools YEAR IV School 21 . 6 . 18 . 14 . 26 . 24 . 4 . Average Score .... 3.97 .... 3-65 .... 3-93 .... 392 .... 4.07 .... 4.37 .... 4-45 School 10 . 2 . I . 7 • 13 . 5 . Average Score .... 4.78 .... 4.80 .... 4-97 .... 5.09 .... 5-50 .... 5.63 J^a3485 38 LATIN IN SECONDARY SCHOOLS study in these schools resulted in an achievement such that at the end of the fourth year the average performance is represented by TABLE 24 LATIN SENTENCE TEST A Record of Impro\t:ment In Terms of the Scores Made by Pupils CLASS AVERAGES Year I . II . Class Averages .... 2.88 .... 3.58 Year III . IV . Class Averages .... 3.98 .... 4.46 1 See footnote, Table 14. ability to translate a sentence of the grade of difificulty of the sixth sentence in Test B. In general the same facts are shown in the results of Latin sen- tence Test A. The first twelve of these sentences are extremely easy, most of them having been selected from the first few pages of beginners' books. Thus all but a few of the sentences of this test are of the grade which pupils are supposed to have in their first year of Latin study and nearly the first dozen are such as they would probably meet in the first three months. In Test A all the pupils of Year I secured an average about equal to the scale value of the fourth sentence. In Year II the gain in ability was such that the average performance about equalled the difficulty of the eleventh sentence. There was enough improve- ment during Year III so that the average performance was slightly above Sentence 12. During the fourth or final year of Latin study the ability of the pupils reached a point represented by Latin of the grade of difficulty represented by the thirteenth sentence. In a word, as far as the results of these sentence tests go, it is clear that the improvement in ability to read the meaning of the simplest Latin is meager out of all proportion to the time, energy, and money devoted to instruction in this subject. CHAPTER VI THE CONNECTED LATIN TEST Nature of the Test The test which was used for a connected Latin test is repro- i 97 113 17 I5-0 1.537 389 98 112 12 10.7 1.843 4-19 99 112 7 6.2 2.281 4.63 100 112 5 4.3 2.546 3.90 loi 112 22 19.6 1.269 3-62 102 112 16 14.3 1.582 3.93 103 112 4 3.5 2.686 5.04 104 106 40 27-7 .464 2.82 105 112 3 2.6 2.881 5.23 106 112 38 33.0 .652 3.00 107 Ill 35 31-5 714 3-07 108 Ill 29 26.1 .949 3.30 109 Ill 2 1.8 3.111 5.46 from the first point. Table 25 gives the number of pupils who attempted the Latin corresponding to each point, the number who had each point correct, the per cent who had each point correct, the P. E. values and the weightings for each point. In tables 29-32 are given the results for all the schools. It should be added that in several schools an unlimited amount of time was given for the test in order that the Latin corresponding to every point in the key might be attempted by some pupils. Thus every point was attempted by more than a hundred pupils. The results from the schools which had unlimited time are not included. The method of scoring finally adopted in the Connected Latin Test was based upon three measures. I. The amount attempted. This was found for each pupil by taking the aggregate of the scale values of the points of the key in the amount of Latin attempted by the pupil in the time allowed. II. The amount correct. This was found also by taking the ag- gregate of the scale values of the points which were correct. This could be done very rapidly with an adding machine. III. Comprehension. This was found by determining the per cent correct of the amount attempted. In order to get this for any pupil it was necessary only to divide his mark for the amount correct by his mark for amount attempted. 50 LATIN IN SECONDARY SCHOOLS It is very desirable to have a single measure with which to ex- press a pupil's total efficiency or that of an entire class. It is of so great importance to have such a measure of this kind that some slight sacrifice in statistical accuracy would be warranted if such a measure could be found. Thus far, however, no satisfactory method of finding a single coefficient of efficiency has been discovered. TABLE 26 ABILITY TO APPREHEND THE MEANING OF LATIN Connected Latin Test Distribution Table for Amount Attempted In Terms of Number of Points of the Key Amount Number of Pupils Attempted Year II Year III Year IV 5 •• 5 I- 5 6-10 2 2 I II-I5 23 S 4 16-20 32 7 8 21-25 54 12 9 26-30 125 57 38 31-35 68 44 30 36-40 121 76 68 41-45 41 41 43 46-50 20 18 15 51-55 14 15 16 56-60 13 15 16 61-65 3 8 8 66-70 3 71-75 I 76-80 81-85 I 86-90 2 2 91-95 I 96-100 It has been necessary, therefore, to express the pupils' general efficiency in terms of two marks — amount correct and comprehen- sion. The amount correct is believed to be a rather good measure of the pupils' ability. When there is added to this the mark for com- prehension, we have an accurate measure of ability, although it is somewhat cumbersome by being expressed by two separate marks. CONNECTED LATIN TEST 51 The amount correct indicates what the pupil can do correctly in the specified time. The comprehension indicates his degree of accuracy or his freedom from error in rendering Latin into English. A high score for comprehension would indicate that the pupil had correct about all that he did. If two pupils had the same mark for amount correct, but one had a high mark for comprehension and the other TABIiE 27 ABILITY TO APPREHEND THE MEANING OF LATIN Connected Latin Test Distribution Table for Amount Qjrrect In Terms of Number of Points of the Key Amount Correct o . I- 5 • 6-10 . II-I5 • 16-20 . 21-25 . 26-30 . 31-35 • 36-40 . 41-45 • 46-So . 51-55 • 56-60 . 61-65 • 66-70 . 71-75 . 76-80 . 81-85 . 86-90 . 91-95 • 96-100 Number of Pupils Year II Year III Year IV 21 8 5 71 13 3 98 24 21 90 50 19 85 67 57 90 63 44 37 46 45 20 29 41 8 6 16 2 4 7 I • • 2 I I I had a low mark, it would indicate that the one who had the low mark for comprehension attempted a good deal more Latin than the other and consequently was inaccurate. 52 LATIN IN SECONDARY SCHOOLS General Results of Connected Latin Test In Tables 29-32 are shown the main facts for abihty to appre- hend the meaning of Latin. Table 32 gives an idea of the results in each year in amount attempted, amount correct and comprehen- sion. TABLE 28 ABILITY TO APPREHEND THE MEANING OF LATIN Connected Latin Test Distribution Table for Q)mprehension In Terms of Number of Points of the Key Number of Pupils Comprehension Year II Year III YearlV o 25 7 5 .01- .05 25 6 I .06- .10 24 12 2 .11- .15 43 II 7 .16- .20 33 13 7 .21- .25 36 21 9 .26- .30 35 27 8 .31- -35 35 22 16 .36- .40 35 19 20 .41- 45 43 23 22 .46- .50 42 21 20 .51- .55 35 20 26 .56- .60 28 24 18 .61- .65 22 19 15 .66- .70 13 19 19 .71.- .75 18 14 16 .76- .80 7 10 10 .81- .85 13 8 16 .86- .90 9 8 14 •91- -95 2 6 4 .96-1.00 2 2 5 In the tables referred to, the distribution of amount attempted, amount correct and comprehension is given in terms of number of points of the key. In Table 32 are given the class averages based on the individual scores of all the pupils in each year. The average score for all the pupils of Year II for amount attempted is equiva- lent to the aggregate scale values of the first thirty-seven points of the test, of which the amount correct is equivalent to the aggregate CONNECTED LATIN TEST 53 scale values of the first ten points of the test. This gives a compre- hension of twenty-eight. In other words, these pupils translated correctly less than one-third of the amount which they attempted. For all pupils of Year III, the amount attempted is equivalent to the aggregate scale values of the first forty points of the test, and the TABLE 29 ABILITY TO APPREHEND THE MEANING OF LATIN Connected Latin Test Class Averages In Terms of the Scores Made by Pupils YEAR II Number of Amount Amount Compre- ^"^ °°^ Pupils Attempted Correct hension 9 IS 65.15 7.65 -12 26 16 79-30 7-67 -09 6 12 77.68 12.28 .16 17 34 4426 12.99 -29 3 S 5378 13-79 -26 23 19 70.69 16.86 .24 4 9 7346 18.21 .25 2 14 75.67 19-02 .25 20 5 59-17 19-33 -33 14 12 61.02 21.99 -36 16 96 54.14 22.11 41 18 20 98.69 22.14 -23 31 42 84.89 22.83 '27 24 10 99-58 23.61 .24 19 II 69.27 26.69 .38 12 23 77.94 27.14 .35 7 II 82.53 27.75 -34 10 15 65.70 28.04 -43 21 19 79-50 3I-81 .40 I 9 76.87 32.79 -43 25 43 83.29 34-21 -41 30 19 74-07 35-38 .48 34 15 65.83 35-43 -54 13 4 103-99 38-00 .36 27 4 56.45 38.42 .68 22 13 78.71 41-64 -53 15 3 88.81 58.04 .65 5 23 84.16 60.86 .72 29 5 1 12.61 73-40 .65 54 LATIN IN SECONDARY SCHOOLS amount correct is equivalent to the first twenty points, which yields a comprehension of forty-five. At the end of the fourth year the amount correct is approximately equivalent to the aggregate scale values of the first twenty-five points of the test. Four years of study have thus developed an average ability to write in English TABLE 30 ABILITY TO APPREHEND THE MEANING OF LATIN Connected Latin Test Class Averages In Terms of the Scores Made by Pupils YEAR III Number of Amount Amount Compre- ^'^^''"^ Pupils Attempted Correct hension 20 7 32.26 11.38 .35 9 8 72.10 19.91 .28 10 14 60.55 22.12 .37 26 9 74.40 23.40 .31 • 2 II 119.67 25.38 .21 21 21 89.14 26.70 .30 31 22 58.56 28.33 ul8 14 6 94.96 29,63 .31 3 6 42.50 30.22 .71 18 20 116.42 32.20 .28 16 40 S8.i8 3523 .61 22 6 76.72 35.48 46 4 7 96.60 37.38 .39 25 26 80.82 38.11 .48 27 I 92.18 3909 -42 12 13 82.56 40.56 .49 23 13 96.67 43.94 -45 17 27 7382 43-99 .60 6 4 115.89 44-51 .38 29 2 79.33 44.73 -56 30 5 77.25 45.5s .59 19 8 105.46 46.60 .44 I 2 85.15 48.12 .55 1^ I 56.67 52.84 .92 7 14 164.95 55.28 .34 5 10 83.00 56.70 .68 34 3 II 1.83 56.78 .51 24 4 10357 57.92 .56 13 2 108.91 67.02 .62 CONNECTED LATIN TEST 55 in fifteen minutes the thought of a Httle less than fifteen hnes of Latin of the difficulty of that represented in this test, with about eight and a half lines correct. Improvement in Two Years We may next consider the improvement made from year to year. It would appear from Table 32 that from the end of the second year TABLE 31 ABILITY TO APPREHEND THE MEANING OF LATIN Connected Latin Test Class Averages In Terms of the Scores Made by Pupils YEAR IV Number of Attempted Amount Compre- ^*^^°°^ PupUs Amount Correct hension 27 4 40.54 21.58 .53 9 6 45-22 21.74 48 3 3 95-i6 23.65 .25 26 7 61.93 27.63 45 18 19 120.15 3138 -26 16 29 52.98 3332 .63 31 17 55.18 34-o6 .62 4 5 113.36 37-47 -33 20 5 76.68 38.92 .51 21 9 118.70 42.73 -36 12 8 99.71 43-03 .43 2 13 90.71 43.24 -48 14 II 84.18 44.73 -53 2Z 5 103.53 45.34 -44 6 4 108.92 45.62 .42 24 7 106.05 49.30 .46 19 5 107.18 49.50 .46 22 8 79.61 51.57 -65 17 2,7 93.24 52.21 .56 10 8 78.84 53.09 -67 25 7 88.65 54.39 -61 30 7 83.88 58.71 -70 5 12 87.29 60.72 .70 34 I 75.34 63.20 .84 15 5 95-94 65.12 .68 7 9 121.86 70.86 .58 I 5 88.49 70.89 .80 29 I 77.38 73.51 .95 13 4 129.77 85.23 .66 56 LATIN IN SECONDARY SCHOOLS to the end of the third year, there is some improvement. The pupils of the third year attempted somewhat more Latin, and had a some- what larger amount correct. The improvement in comprehension is represented by an increase from twenty-eight to forty-five. During the fourth year there seems to be only a very slight increase in ability. In terms of comprehension this means that fourth year TABLE 32 CONNECTED LATIN TEST Record of Improvement In Terms of the Scores Made by Pupils CLASS AVERAGES! :xz Yg^j. Amount Amount Compre- Attempted Correct henslon n 79.32 22.13 .28 HI 90.07 40.67 .45 IV 94-9-2 49.00 .52 1 See footnote, Table 14. pupils can get correct twenty-four per cent more of what they write than can second year pupils. This seems hardly adequate as a result of two years of intensive study of Latin. Conclusion Concerning Ability to Apprehend the Meaning of Latin From these facts one conclusion is inevitable. These secondary schools, which are well representative of all such schools in the State, and probably of secondary schools in general everywhere, are succeeding in developing, as a result of four years' work, a little more than fifty per cent accuracy in expressing Latin in English. In individual schools it falls as low as 25, 26, 33 and 36 per cent. As a result of so great inaccuracy it naturally follows that the amount of the thought of easy narrative Latin which pupils who have studied the subject four years are able to express correctly in Engli.sh in a given time is extraordinarily meager. Here is defi- nite objective evidence that high school pupils in general, in this State at least, under our present methods of teaching, are incapable of rapid, intelligent apprehension of the thought of ordinary easy Latin. In view of these facts, the teaching of Latin as now con- ducted in these secondary schools must be regarded as highly unsat- isfactory. CHAPTER VII LATIN GRAMMAR TEST The Latin Grammar Test measured the pupils' ability to describe and classify Latin constructions. The Test Directions to pupils: In the sentences given below — (a) Give the names of the constructions represented by the words in italics. (b) Explain in each case the reason for the use of the construction, i. e., give in your own words the rule governing it. Translations of sentences are supplied. EXAAIPLE:— Galba agricola agrum habet. (Galba, the farmer, has a field.) (a) Construction: noun in the appositive case agreeing with Galba. (b) Reason: an appositive agrees in case with the noun which it limits. 1. Multis interfectis,^6f}pidum expugna.\'mius. (Although many had been killed, we took the town by storm.) (a) Construction: (b) Reason : 2. Multos annos bellum gesserunt. (They have waged war for many years.) (a) Construction: (b) Reason : 3. Agricolae boni equis f«iiHeiilUni dabunt. (The good farmers will give grain to the horses.) (a) Construction: (b) Reason: 58 LATIN IN SECONDARY SCHOOLS 4. Viri constantia magna pugnant. (The men fight with great steadfastness.) (a) Construction: (b) Reason: v^ 5. Dixit Belgas unam partem incolere. C*^^^*^^^ " u^VclU^ y (He said that the Belgians inhabit one part.) (a) Construction: (b) Reason: 6. Castra consuluni a nostris paucis diebus capientur. (The camp of the consuls will be taken by our men in a few days.) (a) Construction: (b) Reason: 7. Viro erat pulcherrima domus. (The man had a very beautiful house.) (a) Construction : (b) Reason: .N 8. Galli copiis Romanis inimici erant. {C>cx\\'^\ ■ ^ "^^j__>' (The Gauls were hostile to the Roman forces.) (a) Construction ; (b) Reason: . .Q'.'r'.-'' . . .... 9. Die mihi q«et-milites smt ni illis castris. (Tell nie how many soldiers are in that camp.) (a) Construction : (b) Reason : 10. Viros subsidio exercitui misit. (He sent themen as an aid to the army.) (a) Construction: (b) Reason: 11. Legatio Roma venit quae voluntatem regis cognoscat. (The embassy came from Rome to learn the wish of the king.) (a) Construction: (b) Reason : 12. Sta^inr imperator iussit nuntios quam celerrimos litteras Romam portare. '.^>" '.^ - 1 v ' - n ^ f ''■'■i' ^' *- ^ '^ ^ (The commander immediately ordered the messengers to take the letters to Rome as quickly as possible.) (a) Construction : (b) Reason : LATIN GRAMMAR TEST 59 13. Te tuo loco demovere potuerunt. (They were able to remove you from your place.) (a) Construction: (b) Reason: 14. Galli Romanes magnitudine corporum superant. (The Gauls surpass the Romans in size of body.) (a) Construction: (b) Reason: 15. Ne moremur in urbe totam noctem. (Let us not delay all night in the town.) (a) Construction: (b) Reason: 16. Non multi erant qui sine ullo vulnere cffugerent. (There were not many who escaped without any wound.) (a) Construction: (b) Reason: 17. Utinam ille omnis secum copias eduxisset. (Would that he had led forth all his forces with him.) (a) Construction : (b) Reason: ^. , -^.^ 18. Nisi in6pia telorum fuisset oppugnatione-m-sustinuissent. (If there had not been a lack of weapons, they would have with- stood the siege.) (a) Construction: (b) Reason: 19. Hostes in silvas fugerunt ut a nostris militibus non viderentur. (The enemy fled into the forest so that they were not seen by our soldiers.) (a) Construction: (b) Reason: ^^.^ 20. Cum legio in proelium fortiter isset, hostibus- non diutissime restitit. .^'-c '.\ l'-l f • (Although the legion had gone bravely into battle, it did not very long restrain the enemy.) (a) Construction: (b) Reason: 6o LATIN IN SECONDARY SCHOOLS TAIJLE 33 ABILITY IN LATIX GRAMMAR Distribution Table for Number of Constructions Correct Number of Constructions I 2 3 4 5 6 7 8 9 10 II 12 13 14 15 i6 17 i8 19 20 Yearl • 254 . ii6 73 6o 41 • 32 22 26 21 29 i6 10 8 4 2 I Number of Pupils Tested 7^5 Median Number of Constructions Correct 1.892 25 Percentile 704 75 Percentile 4.81 1 Quartile 2.054 Number of Pupils Year II 116 64 45 54 50 40 32 41 26 31 22 21 12 II 6 7 3 4 4 2 591 arlll Year IV 18 9 15 20 21 14 31 14 31 26 21 17 30 17 28 26 26 12 24 26 22 12 19 19 17 13 16 19 20 II 10 10 8 12 4 M I ID 2 ■ 6 364 304 4-330 7-536 8.750 1.496 4- 1 94 4-731 8.481 11.316 13.158 3-493 3.561 4.214 LATIN GRAMMAR TEST 6i TABLE 34 ABILITY IN LATIN GRAMMAR Number of Pupils in Each Year Who Answered Each G^nstruction Correctly Construction Year I Year II Year III Year IV 1 207 366 300 231 2 217 287 258 219 3 233 229 191 181 4 204 243 206 170 5 136 260 230 157 6 100 196 161 137 7 124 126 165 156 8 127 113 109 188 9 8s 128 158 157 10 104 175 142 103 11 74 145 108 98 12 66 135 107 100 13 57 97 no 113 14 82 75 81 84 15 58 42 70 95 16 14 46 97 91 17 20 36 73 112 18 13 25 105 127 19 46 70 65 41 20 38 29 50 44 62 LATIN IN SECONDARY SCHOOLS TABLE 35 ABILITY IN LATIN GRAMMAR Per Cent of Pupils in Each Year Who Answered Each Construction Correctly Construction Year I Year II Year III Year IV 1 29.0 61.8 82.4 76.0 2 30.3 48.5 70.9 72.0 3 32.6 38.7 52.5 59-5 4 28.S 41. 1 56.6 55.9 5 19.0 44-0 63.2 51.6 6 14-0 33-1 44-2 45-1 7 17-6 21.3 45.3 51.3 8 17.8 19.1 29.9 61.8 9 11.9 21.7 43.4 SI.6 10 14.5 29.6 39.0 33.9 11 10.3 24.5 29.7 32.2 12 9.2 22.8 29.4 32.9 13 8.0 16.4 30.2 37.2 14 11.5 12.7 22.3 27.6 15 8.1 7.1 19.2 31.3 16 2.0 7.8 26.6 29.9 17 2.8 6.1 20.1 36.8 18 1.8 4.2 28.8 41.8 19 6.4 1 1.8 17.9 13.5 20 5-3 4.9 137 145 LATIN GRAMMAR TEST 63 TABLE 36 ABILITY IN LATIN GRAMMAR P. E. Equivalents of the Difference Between Fifty Per Cent and the Per Cent in Each Year Who Answered Each Construction Correctly Construction Year I Year II Year III Year IV 1 0.820 — 0.445 — 1-380 — 1.047 2 0.765 0.056 — 0.816 — 0.864 3 0.669 0.422 — 0.093 —0.357 4 0.842 0.334 — 0.246 — 0.220 5 1.302 0.224 — 0.500 — 0.059 6 1.602 0.648 0.216 0.183 7 1.380 1. 181 0.17s — 0.048 8 1.368 1.296 0.782 — 0.445 9 1.749 i-i6o 0.246 — 0.059 10 1569 0.795 0.414 0.616 11 1.875 I-024 0.790 0.685 12 1.971 1. 105 0.803 0.656 13 2.083 1450 0.769 0.484 14 1.780 1.692 1. 130 0.882 15 2.074 2.177 1.291 0.723 16 3.044 2.103 0.927 0.782 17 2.834 2.293 1-243 0.500 18 3.11 1 2.562 0.829 0.307 19 2.257 1.757 1.363 1.636 20 2.397 2.453 1.622 1.569 TABLE 37 ABILITY IN LATIN GRAMMAR Final Scale Values of Constructions Construction I 2 3 4 5 6 7 8 9 10 Scale "Value ... I. II 1.37 1-75 1.76 1.80 2.24 2.25 2.31 2.33 2.46 Construction II 12 13 14 15 16 17 18 19 20 Scale Value . . . 2.76 . .. 2.78 -.. 2.79 - ■ 3-01 -.. 3-15 - .. 3-21 -•- 3-22 -•- 329 ■ •• 3-34 . .. 3.60 64 LATIN IN SECONDARY SCHOOLS TABLE 38 ABILITY IN LATIN GRAMMAR Average Scores by Schools YEAR I School 17 .. 27 •• I . , 14 .. 9 •• 6 .. 23 .. 32 ■■ 3 •• 19 .. 30 .. 12 . . 31 •• 7 •• 26 .. Average Score 07 19 30 • ■ • -37 ... .36 28 ... .46 ... .50 ... .56 ... .50 ... .71 ... .87 ... .87 ... .91 ... .90 School 10 . . 4 .. 35 •• 16 .. 21 . . 20 . . 24 .. 13 •• 11 . . 22 29 .. 33 •• 34 •• 25 •• 5 •• Average Score 1.09 .... 1. 16 . ... 1. 14 ... 1.20 60 1.20 . .. 1.25 . .. 1.37 1. 12 1.69 ... 1.72 ... I .76 2.24 . . . 2.24 ••■ 2.33 TABLE 39 ABILITY IN LATIN GRAMMAR Average Scores by Schools YEAR II School Average Score 14 59 27 85 18 92 23 1.05 7 109 6 1. 16 17 I.I7 16 1.22 9 1. 18 26 1.29 2 1.27 3 1.49 4 1.64 32 1.63 31 1.63 I 1.66 13 1.72 School Average Score 20 1. 81 10 1.73 33 1.87 35 1-83 30 1.86 25 1.85 19 1.98 22 2.07 11 2.09 21 2.17 34 2.29 8 2.28 29 2.29 12 2.3s 24 2.39 5 2.79 15 301 LATIN GRAMMAR TEST 65 TABLE 40 ABILITY IN LATIN GRAMMAR Average Scores by Schools YEAR III School 14 .. 17 •• 18 .. 20 . . 6 .. 27 .. 23 .. 7 ■• 33 ■■ 3 • 2 .. I . . 4 •• 32 .. 25 •■ 21 . . 9 •• Average Score 23 1.46 .... 1.66 .... 1.65 — 1.64 . ... I.7S 1.72 .... 1.66 , ... 1.80 . ... 1.82 1.92 2.00 2.07 2.06 2.10 2.13 . . . . 2.19 School 11 . . 26 .. 35 •• 10 . . 34 •• 12 . . 16 .. 24 .. 13 •• 30 .. 19 .. 15 ■■ 5 • ■ 22 . . 8 .. 29 .. 31 •• Average Score 2.24 . ... 2.28 2.25 . ... 2.34 .... 2.59 2.52 . ... 2.57 2.52 2.70 2.61 2.67 2.67 2.76 2.91 . ... 3.00 . ... 3.01 2.30 School 20 . . 18 .. 6 .. 3 •• 17 .. 26 .. 21 . . 2 . . 9 .. 16 .. 27 .. 32 . 23 • 31 • 24 ., 4 . 8 . TABLE 41 ABILITY IN LATIN GRAMMAR Average Scores by Schools YEAR IV Average Score . ... 3.10 1.23 . ... 1.50 . ... 1.74 , ... I.91 1.96 . . . . 2.03 2.06 . ... 2.08 2.07 , . . . 2.14 2.25 .... 2.18 2.26 2.27 . ... 2.33 2.41 School 7 •• 13 ■■ 10 . . 35 •■ 19 .. I 22 25 30 12 II 34 33 15 5 29 Average Score 2.46 ... 2.66 . ... 2.65 2.46 2.76 2.72 . .. 2.73 , ... 2.74 . ... 2.74 . ... 2.77 . ... 2.82 . ... 2.79 . ... 3.01 .... 2.99 .... 3-15 • .•• 3-34 66 LATIN IN SECONDARY SCHOOLS TABLE 42 ABILITY IN LATIN GRAMMAR Record of Improvement In Terms of the Scores Made by Pupils CLASS AVERAGES! I Year Class Averages I MI 11 1-59 Year Class Averages III 2.13 IV 2.27 1 See footnote, Table 14. Conclusion Concerning Ability in Grammar An examination of the data shows that the Latin Grammar Test was too difficult for first year pupils. There is a very large group who failed to get a single construction correct and a smaller but relatively large group who had but one construction correct. In the second year a considerable number had none right. The test seemed fairly well adapted to the third and fourth year pupils. The futility of the work in formal grammar is well set forth in Tables 38-42. It will be seen by examining Table 38 that seventeen schools failed to get an average score as high as the scale value of the easiest construction. In the second year eighteen schools had an average score which was less than the scale value of the third construction. In the first year no school had a score as high as the scale value of the tenth construction. In the second year the best school had an average score identical with the scale value of the fourteenth construction. The class average of all pupils in the first year was identical with the scale value of the easiest construc- tion. In the second year the class average was slightly larger than the scale value of the second construction. All third year pupils averaged about as high as the value of the sixth construction. In the fourth year the class average is approximately the same as the value of the seventh construction. One can but be surprised at the meagerness of the result, on the whole, in this test. The constructions nearly all come within the range of first year Latin and it would seem as though the majority of pupils in third and fourth years ought to answer them all. The average in none of the four years is half the number of construc- tions. Formal knowledge of construction, it will be remembered, is a thing which these schools make to a large degree the end and aim of Latin instruction, and it seems clear that when less than nine out of twenty is the average for 300 pupils after four years' work, there must be waste in the process. CHAPTER VIII LATIN VOCABULARY TEST Nature of the Test In the Latin Vocabulary Test the same plan was carried out as in the grammar test. The method of scoring was the same as in the case of the grammar test. The words of the vocabulary test were chosen from a list of words found in each of seven Latin begin- ners' books in common use. The test is reproduced below. Latin Vocabulary Test I. semper 14. inopia 27. numquam 40. cotidie 2. proelium 15- prope 28. fio 41. fortis 3- facio 16. dom.us 29. accipio 42. timeo 4. spes 17. fossa 30. post 43. scutum 5- vallum 18. nihil 31- murus 44. scio 6. capio 19. saepe 22. regnum 45. hostis 7. auxilium 20. miles 2>?,- gero 46. ipse 8. supero 21. hiems 34- bene 47- munio 9- pilum 22. salus 35- annus 48. quisquam 10. etiam 23- aliquis 36. acies 49- copia 11. intellego 24. vinco 2,7. pedes 50. pax 12. quaero 25- pervenio 38. periculum 13. iubeo 26. nolo 39- Conor TABLE 43 ABILITY IN LATIN VOCABULARY Distribution Table for Number of Words Correct Number of Words I 2 3 4 5 6 7 Year I Year 8 4 9 I 5 2 4 2 8 I 9 6 2 13 Number of Pupils Year III Year IV 68 LATIN IN SECONDARY SCHOOLS TABLE 43 — Continued Number of Words Year I 8 10 9 12 10 10 u 7 12 8 13 7 M 7 15 8 i6 7 17 8 i8 4 19 6 20 6 21 6 22 7 23 10 24 5 25 7 26 7 27 8 28 II 29 12 30 5 31 6 32 12 33 II 34 13 35 12 36 10 37 8 38 II 39 17 40 15 41 13 42 13 43 12 44 4 45 3 46 5 47 2 48 2 Number of Pupils Year II Year III I I I I I I I 2 I I 2 I 2 2 I 4 5 4 8 10 2 7 8 3 9 9 13 12 9 12 12 II 8 II I 9 Year IV 2 3 I 2 2 4 3 3 3 II 8 ID 9 13 5 16 9 6 4 I 4 7 9 12 9 10 II 5 7 10 2 4 LATIN VOCABULARY TEST 6^ TABLE 43 — Continued Number of Number of Pupils Words Year I Year II Year III Year IV 49 6 4 2 50 6 I 1 Number of Pupils Tested 409 211 117 104 Median Number of Words Correct 27.063 39-3o8 42.500 42.000 25 Percentile 12.156 31-938 40.386 39-444 75 Percentile 37-344 43-875 46.422 4S-OOO Quartile 12.594 5-969 3-Oi8 2.778 TABLE 44 ABILITY IN LATIN VOCABULARY Number of Pupils in Each Year Who Gave the Correct Meaning of Each Word Word Year I 1 287 2 292 3 320 4 321 5 265 6 288 7 286 8 251 9 264 10 290 11 296 12 320 13 270 14 273 15 218 16 253 17 235 18 198 19 253 20 249 21 177 22 169 23 272 24 220 25 239 26 243 Year II Year III Year IV 190 115 83 187 115 81 200 115 83 198 117 83 185 113 83 194 "5 80 194 117 80 190 "3 80 184 no 80 190 115 81 JS3 93 68 198 112 80 195 no 81 192 112 80 18s 115 83 193 108 79 154 TOO 78 181 III 79 184 109 72 175 no 76 180 III 81 164 104 81 149 95 74 170 105 72 171 92 76 160 lOI 73 70 LATIN IN SECONDARY SCHOOLS TABLE 44 — 'Continued Word Year I 27 179 28 203 29 173 30 204 31 182 32 246 33 145 34 188 35 162 36 162 37 149 38 202 39 331 40 188 41 129 42 134 43 146 44 93 45 71 46 41 47 53 48 44 49 22 50 48 TABLE 45 ABILITY IN LATIN VOCABULARY Per Cent of Pupils in Each Year Who Gave the Correct Meaning for Each Word Word Year I 1 70.2 2 71-4 3 78.2 4 78.5 5 64.8 6 70.4 7 69.9 8 61.4 9 64.5 10 70.9 Year II Year III Year I 152 III 83 164 108 76 172 lOI 73 157 91 69 170 lOI 82 131 87 73 155 95 70 172 96 64 159 92 73 138 78 52 134 91 67 139 81 55 181 114 80 121 68 55 116 81 61 125 72 62 143 77 60 81 88 68 100 71 54 103 60 64 82 68 49 86 45 18 62 61 44 198 45 45 Year II Year III Year IV 90.5 98.3 79.8 89.0 98.3 77.9 95-2 98.3 79.8 94-3 1 00.0 79-8 88.1 96.6 78.8 93-2 98.3 76.9 92.3 1 00.0 76.9 90.5 96.6 76.9 87.6 94.0 76.9 90.5 98.3 77-9 LATIN VOCABULARY TEST 71 TABLE 45 — Continued Word Year I 11 72.4 12 78.2 13 66.0 14 66.7 15 53-3 16 61.9 17 57.5 18 48.4 ig 61.9 20 60.9 21 43.3 22 413 23 66.5 24 53-8 25 58.4 26 59-4 27 43-8 28 49-6 29 42.3 30 49-9 31 44-5 32 60.1 33 35-5 34 46.0 35 39.6 36 39-6 37 36.4 38 49-4 39 80.9 40 45-9 41 3I-S 42 32.8 43 37.5 44 22.7 45 174 46 lO.O 47 13-0 48 10.8 49 • 5-4 50 1 1.7 Year II Year III Year IV 72.9 79-5 65.4 94.3 95-7 76.9 92.9 94.0 77-9 91.4 95-7 76.9 88.1 98.3 79.8 91.9 92.3 76.0 73-3 85.5 750 86.2 94-9 76.0 87.6 93-2 69.2 83.3 94-0 73-1 85.7 94-9 77.9 78.1 88.9 77-9 71.0 81.2 71.2 81.0 89.7 69.2 81.4 78.6 73-i 76.2 86.3 70.2 72.4 94-9 79.8 78.1 92.3 73-1 81.9 86.3 70.2 74.8 77.8 66.3 81.0 86.3 78.8 62.4 74-4 70.2 73.8 81.2 67.3 81.9 82.1 61.5 75-7 78.6 70.2 65.7 66.7 50.0 63.8 77.8 64.4 66.1 69.2 52.9 86.2 97-4 76.9 57.6 58.1 52.9 55-2 69.2 58.7 59-5 61.5 59.6 68.1 65.8 57-7 38.6 75-2 65-4 47.6 60.7 51-9 49.0 51-3 61.5 39-0 38.1 47.1 41.0 38.5 17.3 29-5 52.1 42.3 94-3 28.S 43-3 72 LATIN IN SECONDARY SCHOOLS TABLE 46 ABILITY IN LATIN VOCABULARY P. E. Equivalents of Differences Between Fifty Per Cent and the Per Cent in Each Year Who Gave the Correct Meaning for Each Word Word Year I Year II Year III Year IV 1 —0.786 —1.944 —3-146 —1.238 2 —0.838 — 1.819 —3.146 —1. 140 3 —1. 155 —2.468 —3.146 —1.238 4 —1. 170 —2.344 —1.238 5 —0.563 —1.749 —2.706 —0.786 6 —0.795 —2.211 —3.146 — 1.091 7 —0.773 — 2.114 — 1.091 8 —0.430 —1.944 —2.706 — 1.091 9 —0.551 -1.713 —2.305 — 1.091 10 — 0.816 — 1.944 — 3.146 — 1.140 11 — 0.882 — 0.904 — 1.222 — 0.588 12 — I.I55 —2.344 —2.546 — 1.091 13 —0.612 —2.177 —2.305 —1. 140 14 — 0.640 — 2.026 — 2.546 — 1.091 15 —0.123 —1.749 —3.146 —1.238 16 —0.449 —2.074 — 2.114 —1047 17 — 0.280 — 0.922 — 1.569 — i.ooo 18 0.059 — i.6i6 — 2.425 — 1.047 19 —0.449 —1.713 — 2.211 —0.744 20 —0.410 —1.432 —2.305 —0.913 21 0.250 —1.582 —2.425 —1. 140 22 0.326 —1. 150 —1.811 —1149 23 —0.632 — 0.820 — I-3I3 — 0.829 24 — 0.141 —1.302 —1.875 —0.744 25 —0.315 —1.324 —1. 176 —0.913 26 —0.353 -1.057 —1.622 —0.786 27 0.231 —0.882 —2.425 —1.238 28 0.015 —1. 150 —2. 1 14 —0.913 29 0.288 — 1.351 —1.622 —0.786 30 0.004 — 0.991 — 1.13s — 0.624 31 0.205 —1.302 — 0.624 —1.186 32 —0.380 —0.468 —0.972 —0.786 33 0.551 —0.945 -1.313 —0.665 34 0.149 —1.351 —1.363 —0.434 35 0.391 —1033 —1. 176 —0.786 36 0.391 — 0.600 — 0.640 37 0516 —0.523 —1.135 —0.547 38 0.022 — 0.616 — 0.744 — 0.108 39 —1.296 — 1.616 —2.881 — 1.091 LATIN VOCABULARY TEST 73 TABLE 46 — Continued Word Year I 40 0.041 41 0/M 42 0.660 43 0.543 44 I. no 45 I-93I 46 1.900 47 1-670 48 1.835 49 2-384 50 1.765 Year II .284 .194 .357 .698 0.430 0.089 0.037 0.414 0.337 0.799 ■2.344 Year III -303 5.744 '-434 1.603 — 1.009 3.403 ).048 '.303 0.434 -0.078 0.434 Year IV .108 1.322 -0.360 3.288 3.588 3.071 ».434 0.108 O.IOO 0.288 0.250 Word 1 . 2 . 3 - 4 • 5 • 6 . 7 • 8 . 9 • 10 . 11 . 12 . 13 - 14 . 15 - 16 . 17 • 18 . 19 . 20 . 21 . 22 . 23 • 24 . 25 - TABLE 47 ABILITY IN LATIN VOCABULARY Final Scale Values Scale Value . . . 2.70 ... 2.73 ... 2.85 ... 2.86 ... 2.87 . .. 2.88 . .. 2.89 . . . 2.90 . .. 2.91 . . . 2.92 . .. 2.93 • • . 2.94 . .. 2.97 . .. 2.98 ... 308 . .. 3.12 . .. 3-23 .-• 3.25 . .. 3.27 ... 3.28 .. 3.32 . .. 3.33 . .. 3.34 ... 3-35 . .. 3.36 Word 26 . 27 . 28 . 29 • 30 . 31 - 32 ■ 33 - 34 • 35 • 36 . 37 - 38 . 39 • 40 . 41 . 42 . 43 . 44 - 45 • 46 . 47 - 48 . 49 - 50 . Scale Value -.. 3-37 ■ ■■ 3-45 . .. 3-50 ... 3.51 ... 3.55 ... 3.57 ... 3.59 ... 3.60 . .. 3.63 ... 3.68 ... 3-72 ... 3.86 ... 3.87 ... 4.03 ... 407 . .. 4.10 . .. 4-11 . . . 4.22 . .. 4.31 . .. 4.47 • •- 4-52 . .. 4.68 ... 484 . .. 501 ... 5.07 74 LATm IN SECONDARY SCHOOLS TABLE 48 ABILITY IN LATIN VOCABULARY Average Scores by Schools 1)1 Terms of the Scores Made by Pupils YEAR I School Average Score 26 2.99 6 316 18 316 2 3-33 33 3-38 14 3-51 7 3.20 35 307 School Average Score 12 2.65 I 365 21 4.04 4 3-82 10 3-65 17 3-73 13 3.78 5 402 TABLE 49 ABILITY IN LATIN VOCABULARY Average Scores by Schools In Terms of the Scores Made by Pupils YEAR II School 26 ., 4 6 10 14 2 35 7 Average Score . ... 3.23 . ... 3.65 . ... 3-7^ .... 3-44 . ... 3-35 . ... 407 2.90 .... 389 School 18 .. 21 . . 17 .. 13 .. 33 .. I . . 5 .- Average Score , ... 3-97 .... 3.75 4.01 . .. 4.1 1 ••. 3-95 ... 4-52 . .. 4-73 TABLE 50 ABILITY IN LATIN VOCABULARY Average Scores by Schools In Terms of the Scores Made by Pupils YEAR III School 21 . . 35 .. 14 .. 6 .. 18 .. 4 .. Average Score , ... 4.21 . .. 3.86 ... 4.16 . .. 4-05 . .. 4.14 . . . 4.20 School 10 . . 7 .. 26 .. 1 . . 2 . . .T Average Score . ... 4-25 . ... 4.49 4.02 . ... 4.47 . ... 4.14 . ... 4.69 LATIN VOCABULARY TEST 75 TABLE 51 ABILITY IN LATIN VOCABULARY Average Scores by Schools In Terms of the Scores Made by Pupils YEAR IV School 35 •• 6 .. i8 .. 21 . . 14 .. 26 .. 4 •• Average Score 4.06 . .. 3-92 ••• 3.97 . . . 4.02 . . . 4.02 . . . 4.12 . . . 4.21 School 10 . . 33 ■■ 1 . . 7 •■ 2 . . 5 •• 13 •• Average Score 4.10 4.20 . .. 441 .... 4-15 ... 4.48 . .. 456 ... 4-23 TABLE 52 ABILITY IN LATIN VOCABULARY Record of Improvement In Terms of the Scores Made by Pupils CLASS AVERAGES! Year I . II . Class Averages • ••. 3-39 .... 3.89 Tear III . IV . Class Averages .... 4.24 .... 4-17 1 See footnote, Table 14. Conclusion Concerning Knowledge of Vocabulary With reference to knowledge of vocabulary the result seems to be a rather satisfactory one. The efficiency appears to be reason- ably high in each year, and there is an improvement from year to year. The test appears to be neither too hard nor too difficult in any year. The class average for Year I is a little higher than the scale value of the twenty-sixth word. This would appear to be a reason- ably satisfactory result. The second year class had an average as high as the thirty-ninth word, while the third year class had an average nearly equal to the forty-third word. The fourth year class did not average as high as this. The failure in ability to interpret connected Latin or to trans- late Latin sentences apparently is not due to lack of knowledge of easy Latin words. CHAPTER IX TIME DEVOTED TO STUDY OF LATIN Variations in Time Table 53 gives in tabular form the facts concerning the amount of time devoted to the study of Latin by the thirty-five schools included in this investigation. A glance at the table shows that the school which uses the largest time allotment spends on the aver- age more than three times as much time per week throughout four years as does the school with the lowest time allotment. In the first year, School 2 uses less than a third of the total amount of time per week which is used by School 13. The time allotments of these two schools Ao not differ to such a great extent in the mat- ter of class work throughout the four years, although the time allotment of School 13 is somewhat greater, but in the amount of time spent in the study of Latin outside of class the difference is marked. In the first year, for example, the pupils of School 13 devote five times as much time to Latin study outside of class. Do these large time allotments found in the case of some schools pro- duce results commensurate with the time? Or is there waste of effort here? Specific discussion of the efficiency of the work of various schools in relation to their time allotments will be reserved for a later chapter. Norms of Current Practice It is interesting to have before us an array of data like those in Table 53. These are probably well representative of what would be found if the same facts had been collected from all of the schools of the State. With this table before him the school administrator may know how his school compares with current practice in this respect. lie will thus have a valuable aid in the administration of his school. He will be able to see at a glance that the middle time allotments in each half are 418 and 565. By comparing the time spent on Latin in his own school, he will know whether he is spend- ing as much or more time than other schools. By studying the whole problem of time allotments and efficiency, as treated in Chap- ter XIII, he will be able to discover whether the time allotments in his school are justified or not. TIME DEVOTED TO LATIN 77 ° o J< >< u 1) •T3 CD 1> s O rt < t> J I) C o o _ ^ O O OS 01 -^ ^ <-0 IT) o g g o _.o \0 00 Tj- CO rf O lO o •oOoociiooo»ooooioOiomioO"^Ovoo lOO . V O l-H G ° O o J3 O0000000"^0"^0000000l000l000l000l^)lou^00"^"^0 ^O O OsioO moscs) mr^ioioioosO O O t^c^ t^(N mo n -fo cs) mt^o csi i^t^o 0< l-Trro-^-rr-rf-^-rfTfTfrfTi-rfT^iOirjiriioiOiomiO iriso SO O SO SO lOso sO iO\0 00 1-. o 11 ^ 4J en C M OOOOiriOOO Tj-O OO M QtOI^ "OOOOOOOOOOOOu-jioQOOOOioioiriO"^ >-iio'i-iommo\OOOOoio)Oo>oc^orot^r>.u^io rfTj-Tl--^m':1-'^iOu^i/>ir)ioiOiOiOii^'^ trjso u^ lO lO loso lO O O O -00 -to lO trjOO OOOOOOiomOOOOOOioOOOuo'iOtOOOOu^OiOGioOOO rrmr^OsOOl^OOsiOTfir>ioioi-iOOO(^ •jmoimOLOi-iOOOCsf^O'^iO'. _ N fOfOf^oo-^fOt~OCO<~0'^"*'*'^'^'^iOiO-*iO'tiOU-}iOTl- ■^so Tj- ro lO "^so -^SO 00 88 O O O o o< o in o r- 0) IN 0) ° O Q Q lO Q 00 O »0 C< N M c^ CS CS c^OQ"^OOOQOO"^00 inoc^oooooo" ~- ?1 _ _ ^OOOOOOOu-lOOQtoO _ _ _ _ _ r^OsoOmmiou^ocsioot-viOTfot^O f^rOCOfO'^fOrofOfOrO'^Tf'^C'O-^'^rO-^OlsO fOSO a! iO0O>o0ir5OO0Q"0OinO00000O0io0O''^iO rOiOirjMuirMOOOOr^OTj-OiOOOioO roOO t^ O O f^ 1^ O o so 3 >, o 2 I ? 13 — . ■" ° h-l o o en O Q O OJ O lO ►^ !N M 8 m lo lo O O (M t>. -H(s)ir>'^rOOOOOOOOOO(NfOOTl-Ot^Ln r^OO O O Q O m O Q O O O O (N fO O Tj- O U-) O lO O O t^ LO r^CO O O O 01 U-) ID _ l-H ►-. 01 01 iCOiOiOO O QvoQi'^Q O O QiOQO ir5iO0»O0 10010)000010 01 OiOO OlOOS'-' 01 01 01 rorooq roOl cOf^r^ 01 01 ^ 01 01 ro t^ ^ O O CO m o Q ■ "" oosO > O Q O O O L, 01 O uo OS m '-I ,_ 01 K « (S 8Q lO O O 00 00 " * 01 1-1 OS O 1-1 01 ir>0"^0000000iOQ00>o0 OIOOIOSOOOOOIOOOIOOOO t^so "■"OlwOlOlOlOlOli-iOlOl"' lO ir> 01 01 01 01 ID m O _ O t^ o 01 t-i 1—1 1—1 1-1 1-1 01 O lO o o o so O O so 01 1-H 01 c^ 1-1 ^ i- t> n1 OJ O. ^ e o HJ o. c rt J 3 C C E ^ ;-• 14-1 o O S en tj 3 a mOQOQOQioO moioioosioooi •- _2 i-t i-i 01 mOsioO 01 OsOsO -ii-iOlOlOli-ii-i" mioO"^000000>oioOO>oQOmio OlOlOOlOsOOOOOlOlOlOOOlOsOOst^. _ _ _ _ 0101010101i-i010)0101010101010101i-i01t-it-i0101010irO01 8 o 01 8 o 8 a 3 a 1:; w o O 1^ fvi - ^ o o 01 o 'a- o 1-1 01 i-* 01 ID Q lO IT) Q O oi 5 01 0\ OS o 01 01 1-1 >-i 01 loinQinooOOQOioir) OlOlOoiCsOOOOOtOlO) OIOIOIMOIOIOIOIOIOIOI o 8 LO O o m m O o o 8 o Q o 01 OS O Os t^ o o 01 'i- o CM 01 01 01 t-H hH 01 01 01 01 01 01 OOOOOOiOOOOOiOOinQOOOOOLomQQOOQiomQOVpOOO OlOOlOsl^OOlOOOsOOOlOOlOsOOOOOlOloiOOiOONOOsl^OOOOOO „oi — i-iOioioii-ii-ioioioioioi'-ioioioioioioioioioi'^MOiH^i-ioioioioirooi 2 o» Os mso r^ vo 1-1 1-1 f^ trjso oo ■* -i-so A Mi-i i-ifO01', Vol. XI, No. i, Janu- ary, 1920, by W. H. Fletcher, entitled The Translation Method in Latin. METHOD IN LATIN 79 of the English language with absolute correctness without knowing a single fact of grammar. It is claimed that it is thus possible to learn to translate a foreign language without first mastering the science upon which it is based. The first lessons in this method are taught entirely from the blackboard without the use of a book. Many sentences are written and read in each period. Such explanations in relation to form and construction as are essential to the progress of the work are made orally by the teacher. New forms are not introduced in any fixed order, are always met first by the pupils in sentences, are explained by the teacher and assimilated by the pupils by reacting to them again and again as they occur in the translation. The pupils are not required to describe or classify the various forms and construc- tions which they use. To be able to react to them correctly is suffi- cient. The second person of a verb may not be taught for a month after the third person and the pupil may know a certain nominative form for some time before he becomes familiar with the correspond- ing ablative. All of the forms which come within the scope of first year Latin are learned in this way during the year. It is held that making the pupils focally, explictly conscious of a multiplicity of minute particulars of form and syntax at the outset is unfavorable to the best development of power to get thought from the language rapidly and accurately. Pronunciation is learned largely by imita- tion. Each sentence at the beginning is pronounced clearly and dis- tinctly by the teacher and the pupils repeat after him, several times at first, thus, it is claimed, learning pronunciation in a rational man- ner from the lips of the teacher. After a few days of this practice the pupils begin to pronounce for themselves and do so independ- ently thereafter under the guidance and direction of the teacher. Each lesson at first is read irrLatin and translated several times so that each pupil gets abundant practice in reacting correctly to the various forms. Perception Card Drill Perception cards are used as a valuable means of drill on forms and for vocabulary. Words, phrases and idioms are printed on white cards of stiff material four inches wide and nine or ten inches long, and a few minutes of very rapid quick-perception drill are 8o LATIN IN SECONDARY SCHOOLS given on these words each day. In this drill the pupils respond to whatever form appears on the card, with the correct translation. This is considered of especial value in connection with noun and verb forms. Instead of memorization and recitation of paradigms, rapid drill of this kind is given daily and in these exercises the forms never appear in any particular order. Each day a few of the more common of the words, idioms and other short expressions are put on the cards. The number of cards grows throughout the year and after a time a considerable supply accumulates, but as the days go by, the pupils become familiar with the cards which have been in use for a time. Consequently, as fast as they become able to respond automatically to the form on any particular card it is dropped out of those in use. Thus there is a constant process of adding new cards and dropping others. All of them are kept, however, and are reviewed frequently. After a considerable number of cards has accumulated the teacher will use fifty or sixty on one day, as many more the next, and so on, perhaps going over two or three hundred in a week. The drill is made very rapid. It is held that unless the perception is practically instantaneous its value is slight. Chief Emphasis in Learning Forms and Constructions The chief emphasis in this method, however, is placed upon the learning of all forms and constructions in a functual way, i. e., by meeting them again and again incidentally through much transla- tion. No emphasis whatever is put on the formal learning of gram- mar by those schools which use the method as described. Use of Books At the end of about a month's work in this method the class is able to begin the translation of easy Latin from books, and from this point on, the greater part of it is done in this way. There are many excellent beginners' books on the market and some schools have a number of sets and do a large amount of translating during the year. These schools usually also possess a number of sets of books containing easy Latin stories, and very early the class can begin to translate the easiest connected Latin. From this time on a good deal of translation of this kind of material is done. The pupils deal to such an extent, from the first, with connected Latin that, it METHOD IN LATIN 81 is claimed, they get such a feeling for construction that a brief explanation of a new principle, when it is met in translation, often suffices to make its use clear to the pupil. Method in the Upper Years During the years above the first, the same general method is fol- lowed. Whatever grammar is taught is usually presented through prose composition, in which one lesson a week is given. In this, no formal grammar lessons are assigned to be studied and recited. Texts in grammar are used as reference books. Principles of usage are taught orally by the teacher with the use of the blackboard on which to write illustrative sentences. In some schools the work is quite largely carried on in class with little or no study outside. The procedure in teaching a given construction might be something like the following. The pupils would probably have some familiarity with it from having met it in previous translation. The teacher's first step would be to find out by questions just what the class knew about it. This would also serve to recall to their minds all the knowledge which they had concerning it. The teacher would next explain to the class briefly, but clearly and concisely, the essential facts about the principle, illustrating his statements by writing on the board short Latin sentences containing the particular construc- tion in question and having them translated by the pupils. This would be followed by sending the entire class to the board to write sentences containing the construction dictated by the teacher. A good many sentences would be dictated and written in this way in the period. In some schools a few sentences closely connected with the text being translated are written each day, but no more in amount than when one day a week is given to the prose composition. The large emphasis throughout this work is put on the learning of Latin usage functionally and not through a formal study of gram- mar from a text and a book on prose composition, although both are used. The Grammatical Method General Plan of the Method In a part of the schools which were tested, the grammatical method is in use. These schools use a beginner's book containing 82 LATIN IN SECONDARY SCHOOLS a systematic presentation of the elementary facts and principles of Latin grammar. There are grammatical rules illustrated by type sentences to be studied, conjugations and declensions to be learned, Latin sentences to be translated into English and English sentences to be written in Latin. Both the Latin and English sentences are usually based largely on the principles of grammar which the lesson in question is supposed to illustrate. There are usually short con- nected passages for translation throughout the book, but the amount of translation is very limited as compared with what some other schools do. The plan of the book is generally followed somewhat closely. Particular Points in the Grammatical Method The particuluar points to be noted in connection with this m.ethod in the first year are the following : ( i ) There is a very large amount of time devoted to instruction in formal grammar. (2) A good deal of time is spent on the memorization of forms independ- ently of their use in translation. (3) The amount of translation is limited. (4) The amount of time devoted to study of Latin outside of school is large. Method in the Upper Years In the years above the first in a part of the schools, grammar is emphasized daily in connection with the texts which are translated. In some of these schools regular grammar lessons are assigned, studied and recited. Daily attention is given to points of syntax as they are met in the translation. Usually this takes the form of detailed questions about points of grammar and the pupils are re- quired to name, describe and classify the forms and constructions which occur. Grammar is also taught in connection with the weekly exercises in prose composition. In this pupils are usually assigned references to look up, grammar lessons are studied, the examples in the composition text are thoroughly examined and English sen- tences to illustrate the principles of the lesson are written in Latin. Extent of Use of Different Methods A third of the schools teach the grammatical method in about the way we have described it, i. e., in its pure form. Five schools METHOD IN LATIN 83 teach the translation method in some form, — no two of them fol- lowing just the same procedure, — in general according to the prin- ciples which we have described. Three others teach a somewhat close approach to it. The remaining eight schools teach a modifica- tion of the grammatical method, in the direction of the translation method, or a modification of the translation method in the direc- tion of the grammatical method. In these schools a text book is used in the first year, but a good deal less emphasis is put on the formal learning of grammar, a large amount of translation is done from the beginning, perception cards are used for drill, although paradigms are required to be memorized. Above the first year, a large amount of translation is required. Types of Method In general, the different methods in use in the first year in the schools included in the test may be classified under three heads as follows : A. Translation method with little or no systematic formal study of grammar and a large amount of translation. B. (a) Translation method somewhat modified in the direction of the grammatical method, with a small amount of grammar and a good deal of translation. (b) Grammatical method, materially modified in the direc- tion of the translation method, with a good deal of translation. C. Grammatical method with a large amount of systematic study of grammar and a limited amount of translation. The procedure in teaching above the first year in all of the schools may be grouped into the following classes : I. Attention largely given to translation with only the most inci- , dental reference to syntax during the regular class periods. Gram- mar taught once a week in connection with prose composition. Chief emphasis distinctly on learning through use without formal drill. Prose sentences given, intended to illustrate and fix in mind facts of usage which have been met in the translation and learned functionally and perhaps somewhat incidentally. Much translation of Latin into English, a large part of which is at sight. 84 LATIN IN SECONDARY SCHOOLS II. Attention largely given to translation with a limited amount of reference to points of syntax during regular class periods. Thor- ough drill in principles of grammar once a week in connection with prose composition. Considerable translation of Latin. III. Systematic drill in grammar in connection with daily work in translation. Text on grammar used constantly for study and reference. Thorough drill in principles of grammar once a week in connection with prose composition. Types of method are indicated by the designations used in the preceding paragraphs, such as A I, B I and C III. These designa- tions will be used throughout the monograph to indicate methods used by the different schools. For example, the designation A I indicates that the school in question uses the method described under A above in the first year and that under I in the years above the first. CHAPTER XI VALUE OF THE STUDY OF GRAMMAR Statement of the Problem A study of the work of the schools and their reports indicates that there are large variations in the amount of time given to the direct teaching of Latin grammar. The problem involved in the title of this chapter at once suggested itself as one of the important aspects of the investigation. Do the schools which devote a very large amount of time to the direct teaching of grammar secure uni- formly better results in knowledge of construction sufficient to jus- tify the expenditure of this time? If not, there is waste effort in the time devoted to grammar. Amount of Grammar in Different Schools In order to secure some information on this point twenty-four of the schools were arranged in three classes called Groups I, II and III, according to the amount of grammar taught. In the first group were placed those schools which teach the translation method in the first year, with little or no formal study of grammar, or a close approach to it. Five of these schools taught the translation method at the time of the test. Three taught a method which could not be classified as the translation method, but it was a method greatly modified in that direction, and a close approach to it. In all these schools the whole emphasis, as previously pointed out, is put on learning grammar incidentally by means of reacting to grammatical forms again and again in the translation of Latin, and the amount of time devoted to formal teaching of syntax is very much less than in the regular or modified grammatical method. The three schools which use the grammatical method in a much modi- fied form, and with much less emphasis on the direct teaching of grammar and stress on learning grammatical forms and construc- tions through translation, are classed in this group. They are selected as the three schools, in addition to the first five, which in the last four years have put the least emphasis on formal teaching of grammar in the first year. 86 LATIN IN SECONDARY SCHOOLS The schools selected for this comparison of method are schools well known to the writer and in which the results are comparable. Conditions are very similar in these schools. In the case of all of these schools in Group I, in the years above the first, the teaching of grammar is restricted to one period a week and it is all taught in connection with the weekly exercises in prose composition. TABLE 54 COMPARISON OF EFFICIENCY IN KNOWLEDGE OF CONSTRUCTION Class Averages YEAR I Group Method School AI 6 A I 2 A I 19 I A I i8 A I 15 BI 24 BII 29 BII 30 Average of all pupils in this group CII 5 CII 13 CII I CII 10 II CII 3 CII 31 CII 25 CII 16 Average of all pupils in this group cm 12 cm 9 cm 7 III cm 14 cm 4 cm 22 cm 26 cm 27 Average of all pupils in this group Latin Grammar Test .28 47 .50 1-25 1.72 .71 .75 2.33 1.37 •30 1.09 .56 .87 2.24 1.20 1.47 .87 .36 .91 .37 I.16 1.69 •90 .19 .88 VALUE OF GRAMMAR 87 The schools in Group II teach the grammatical method in the first year in a greatly modified form with much less insistence on rigorous drill on grammar than is the case in Group III. They use a beginners' book containing grammar lessons, and considerable time is spent in direct teaching and drill on the principles and facts of grammar, but in a very concrete form and directly in connection with the translation of Latin sentences and connected passages. In TABLE 55 COMPARISON OF EFFICIENCY IN KNOWLEDGE OF CONSTRUCTION Class Averages YEAR II Group Method School Latin Grammar Test A I 6 1. 16 AI 2 1.27 A I 19 1.98 A I 18 .92 I A I 15 3.01 B I 24 2.39 B II 29 2.29 BII 30 1.86 Average of all pupils in this group 1.58 CII 5 2.79 CII 13 1.72 CII I 1.66 II CII 10 1.73 CII 3 1.49 CII 31 1.63 CII 25 1.85 CII 16 1.22 Average of all pupils in this group 1.61 cm 12 2.35 cm 9 1. 18 cm 7 1.09 III cm 14 .59 cm 4 1.64 cm 22 2.07 cm 26 1.29 cm 27 .8s Average of all pupils in this group 1.46 88 LATIN IN SECONDARY SCHOOLS the years above the first the teaching of grammar is restricted to one period a week and is all taught in connection with prose com- position, in which one lesson a week is given. These schools teach somewhat more grammar than those in Group I. The eight schools in Group III teach the grammatical method in the first year in a somewhat extreme and rigorous form. There is also a great deal of attention to formal learning of grammar in con- TABLE 56 COMPARISON OF EFFICIENCY IN KNOWLEDGE OF CONSTRUCTION Class Averages YEAR III Group Method School AI 6 AI 2 A I 19 A I l8 I A I 15 BI 24 BII 29 BII 30 Average of all pupils in the group CII 5 CII 13 CII 1 II CII ID CII 3 CII 31 CII 25 CII i6 Average of all pupils in this group , cm 12 cm 9 cm 7 III cm 14 cm 4 cm 22 cm 26 cm 27 Average of all pupils in this group Latin Grammar Test 1.64 1.92 2.67 1.66 2.67 2.52 301 2.61 2.10 2.76 2.70 2.00 2.34 1.82 2.30 2.10 2.57 2.35 2.52 2.19 1.66 •23 2.07 2.91 2.28 1-75 2.04 VALUE OF GRAMMAR 89 nection with the weekly exercises in prose composition. It is the custom in most of these schools to assign grammar lessons in con- nection with the references in the text on composition, to be studied and recited. In addition to this, persistent daily attention is given to grammar in connection with the translation of the texts studied in the class. TABLE 57 COMPARISON OF EFFICIENCY IN KNOWLEDGE OF CONSTRUCTION Class Averages YEAR IV Group Method School AI 6 AI 2 A I 19 I A I 18 A I 15 BI 24 BII 29 BII 30 Average of all pupils in this group CII 5 CII 13 CII I II CII ID CII 3 CII 31 CII 25 CII 16 Average of all pupils in this group cm 12 cm 9 cm 7 nil cm 4 cm 22 cm 26 cm 27 Average of all pupils in this group 1 By an oversight this test was not given in School 14. Latin Grammar Test 1.50 2.06 2.76 1.23 2.99 2.27 3.34 2.74 2.03 3-15 2.66 2.72 2.65 174 2.26 2.74 2.07 2.38 2.77 2.08 2.46 2.33 2.73 1.96 2.14 2.41 90 LATIN IN SECONDARY SCHOOLS Schools in Group III make reports like the following: (a) Grammar is taught daily in connection with classroom work. (b) Once a week, study of special rules in connection with composition. Twice a week, in second year, assignments in Bennett's Latin Grammar. (c) Constant practice in grammatical principles is had by all of the advanced classes in connection with the text. (d) During some part of each year, there is a review of grammatical forms, regular and irregular. Grammatical constructions are noted con- stantly as they occur in the reading. TABLE 58 COMPARISON OF EFFICIENCY IN KNOWLEDGE OF CONSTRUCTION Summary Group II III I II III I II III YEAR I II III VALUE OF GRAMMAR gx (e) For each of these upper years: Review of declensions and conjuga- tions. Syntax in connection with text read. (f) All classes above freshmen are provided with grammars and use them continuall}^ for reference work. (g) Daily exercises in grammar in connection with the constructions of the text. Tables 54-57 set forth the essential facts. Table 58 gives a summary of them. It is apparent that in the case of the first year in grammar, the schools of Group I give about as good a re- sult as those in Group III. Group II shows a better result than Group III, but it is also better than Group I. In the second year the schools of Group I are about the same as those of both of the other groups in grammar. In the third year the results in Group I are better than the results in Group III, but Group I falls slightly below Group II. In the fourth year there is no very great differ- ence in the results in these groups. In grammar. Group I stands slightly lower than either of the other two groups. In the above statement the results for the first year should not be taken too seriously, for the reason that the standings of two schools in Group I are absent. Conclusions Concerning Time Devoted to Grammar Here seems to be a perfectly clear case concerning the justifia- bility of the large amount of time given to the teaching of grammar in the schools represented in Group III. There is not enough dif- ference anywhere between Groups I, II and III to warrant saying that one has any particular superiority. In other words, pupils have no greater mastery of grammar in those schools which teach the grammatical method in the first year in its extreme form, which have a systematic study of grammar in connection with weekly exer- cises in prose composition, and which give daily attention to syntax, with constant use of a text book on grammar, along with the study of Latin authors. In view of this fact, the large amount of the pupils' time and the teachers' effort given to the direct study of grammar is unjustifiable. Here is an important source of waste in teaching. Our observation of the work of these eight schools convinces us that results in ability to get the thought of the Latin, which are so 92 LATIN IN SECONDARY SCHOOLS meager, would be less so if a shift of emphasis should be made from grammar to translation or reading of Latin. In a good many schools, half the time of the recitation period is often spent in dis- cussing the minutiae of form and syntax, which according to the evidence contributes almost nothing to either ability to translate or read, or knowledge of construction. It is moderate language, then, to say that the waste here is extravagant. Public funds are liter- ally being thrown away in the expenditure of this large amount of time and efifort which results in nothing of value. If we now refer to Table 53, we shall see that of the schools which comprise the lowest half in time, only two are schools of Group III. Only two of these eight schools which devote most time to grammar have a time allotment lower than that of the mid- dle school in Table 53 and none stands in the lowest one-fourth as regards time. It is true that seventy-five per cent of the schools of Group III are in the two highest groups in amount of time devoted to the subject, and one hundred per cent are in the three highest groups as shown by Table 53. We have here a good indication of the location of the waste which takes place in those schools which devote an excessively large amount of time to Latin and secure no better results than other schools which devote much less time to it. CHAPTER XII RELATION BETWEEN ABILITY TO APPREHEND THE MEANING OF LATIN AND KNOWLEDGE OF CONSTRUCTION The question of the relationship of knowledge of construction and the ability to get the thought of Latin was one of the important problems connected with this investigation. Do the schools which are best in knowledge of construction on the part of their pupils rank highest in ability to grasp the meaning of Latin ? The schools in each year were arranged in four groups in order of efficiency in the Connected Latin Test, as shown in Tables 59-61. In Table 62 the group averages are shown by years, and it appears from these that in each of the years there is, in general, a very slight improve- ment in grammar from Group I to Group IV. There seems to be a very clear case in this aspect of the study. The schools of Group IV in each of the three years are conspicu- ously more efficient in ability to apprehend the meaning of Latin. But they are in every case only slightly more efficient in grammar. For example, in the fourth year the difference in ability in grammar between Group I and Group IV is represented by the ability to answer correctly eight more constructions in Group IV than could be done in Group I. Nearly similar statements might be made in most of the groups in each year. Plainly the superior ability in getting the thought of the Latin on the part of the schools in Group IV in each year is not accompanied by marked superiority in gram- mar. Wlien we examine Tables 59-61, it is evident that there is no hard and fast connection between knowledge of grammar and abil- ity to get the thought of Latin. The fact that a school stands high in construction does not necessarily insure a high standing in ability to get the thought from the Latin. Schools with high and low scores in grammar are found in all divisions. 94 LATIN IN SECONDARY SCHOOLS The final conclusion of this whole problem is that a high score in knowledge of construction does not necessarily insure a corre- spondingly great ability to apprehend the thought of the Latin TABLE 59 RELATION BETWEEN ABILITY TO APPREHEND THE MEANING OF LATIN AND KNOWLEDGE OF CONSTRUCTION YEAR II Connected Latin Test o >, 1 Amount Compre- Latin Group School Correct hension Grammar Test 9 7.65 12 1. 18 26 7.67 9 1.29 6 12.28 16 1. 16 I 17 12.99 29 1. 17 3 1379 26 1.49 23 16.86 24 1.05 4 18.21 25 1.64 Average of all pupils i3S4 22 1.20 2 19.02 26 1 .27 20 19.33 33 i-8i 14 21.99 36 .59 II 16 22.11 41 1.22 18 22.14 23 .92 31 22.83 27 1.63 24 23.61 24 2.39 Average of all pupils 23.19 31 1.30 19 26.69 38 1.98 12 27.14 35 2.35 7 27.7s 34 1.09 III 10 28.04 43 1-73 21 31.81 40 2.17 I 32.79 43 1.66 25 34.21 41 1.85 Average of all pupils 30.16 38 1.90 30 35.38 48 1.86 34 35.43 54 2.29 27 38.42 68 .85 IV 22 41.64 S3 2.07 15 58.04 65 3-01 5 60.86 72 2.79 29 7340 6s 2.29 Average of all pupils 51.64 58 2.15 MEANING AND KNOWLEDGE OF CONSTRUCTION 95 rapidly and accurately, although there is some correspondnce, in general, with exceptions in particular cases. Nearly all schools which teach Latin, as previously pointed out, spend about a fifth of their time teaching prose composition, not TABLE 60 RELATION BETWEEN ABILITY TO APPREHEND THE MEANING OF LATIN AND KNOWLEDGE OF CONSTRUCTION YEAR III Connected Latin Test a V, 1 Amount Compre- Latin Group school Correct henslon Grammar Test 20 11.38 36 1.65 9 19.91 28 2.19 10 22.12 :i7 2.34 I 26 23.40 31 2.28 2 25.38 21 1.92 21 26.70 30 2.13 31 28.33 48 2.30 Average of all pupils 20.98 32 2.17 14 29.63 31 .23 3 30.22 71 1.82 18 32.20 28 1.66 II 16 3523 61 2.57 22 35.48 46 2.91 4 37-38 39 2.07 25 38.11 48 2.10 Average of all pupils 32.16 39 .2.14 27 39-09 42 I-7S 12 40.56 49 2.52 23 43-94 45 ^-72 III 17 43-99 60 1.46 6 44-51 38 1.64 29 44-73 56 3-01 30 45-55 59 2.61 Average of all pupils 44.00 48 1.87 19 46.60 44 2.67 I 48-12 55 2.00 15 52.84 92 2.67 IV 7 55-28 34 1-66 5 56.70 68 2.76 34 56.78 51 2.59 24 57-92 56 2.52 Average of all pupils 55-19 54 2.32 96 LATI.y IX SECONDARY SCHOOLS because they expect that their pupils will ever have occasion to write anything in Latin after they get out of school, but in order that they may better understand the constructions of the language and thereby be able to read or translate Latin better. It is very TABLE 61 RELATION BETWEEN ABILITY TO APPREHEND THE MEANING OF LATIN AND KNOWLEDGE OF CONSTRUCTION YEAR IV Connected Latin Test ^ -„ , , Amount Compre- Latin Group School Correct henslon Grammar Test 27 21.58 53 2.14 9 21.74 48 2.08 3 23.65 25 1.74 I 26 27.63 45 1.96 18 31-38 26 1.23 16 33.32 63 2.07 31 34.06 62 2.26 Average of all pupils 28.19 44 i-93 4 37-47 33 2.33 20 38.92 51 3.10 21 42.73 36 2.03 II 12 43.03 43 2.77 2 4324 48 2.06 23 45-34 44 2.18 Average of all pupils 43-00 40 2.29 6 45.62 42 1.50 24 49.30 46 2.27 19 49-50 46 2.76 III 22 51.57 65 2.73 17 52.21 56 1.91 10 53.09 67 2.65 25 54-39 61 2.74 Average of all pupils 51.17 53 2.33 30 58.71 70 2.74 5 60.72 70 3.15 34 63.20 84 2.79 IV 15 65.12 68 2.99 7 70.86 58 2.46 I 70.89 80 2.72 29 73-51 95 3-34 Average of all pupils 65.25 73 2.84 MEANING AND KNOWLEDGE OF CONSTRUCTION 97 clear, from a study of many papers, that a pupil may be so taught and drilled in grammar that he can translate very correctly datives and ablatives and subjunctives, and be able to describe and classify Latin constructions with ability, but yet may not be a good thought- TABLE 62 RELATION BETWEEN ABILITY TO APPREHEND THE MEANING OF LATIN AND KNOWLEDGE OF CONSTRUCTION Summary Connected Latin Test Q Amount Compre- Latin Correct hension Grammar Test YEAR II I 13-54 22 1.20 H 23.19 31 1.30 III 30.16 38 1.90 IV 51.64 58 2.IS YEAR III I 20.98 32 2.17 n 32.16 39 2.14 III 44.00 48 1.87 IV 55.19 54 2.32 YEAR IV I 28.19 44 1.93 II 43.00 40 2.29 III 51-17 53 2.33 IV 65.25 72, 2.84 getter when he is confronted with an easy connected passage of Latin to translate. What, now, shall we say concerning the practice of spending so large an amount of time for three years on prose composition for the purpose of clarifying and fixing principles of grammatical construction? What shall we say concerning the ten- dency in one-third of the schools to spend a considerable amount of time and in another third to devote a very large amount of time to direct teaching of grammar? The facts serve only to strengthen the conviction stated in an earlier chapter that the excessive amount of time and attention given to grammar is an unwise and wasteful practice. CHAPTER XIII ABILITY IN THE FUNDAMENTALS OF LATIN IN RELATION TO TIME DEVOTED TO LATIN STUDY It seems desirable at this point to inquire into the exact relation- ship of the amount of time devoted to Latin and the results secured by the various schools. Arrangement of Schools In setting forth the facts in connection with this problem, the schools in each year were arranged in three groups. The schools were arranged in order of the time devoted to the subject. The rela- tionship of time allotments and ability to get the thought of the Latin may be seen by inspection of Tables 63-67. Conclusion Concerning Time Devoted to Latin AND Efficiency It appears from the tables in which the results are set forth that there are somewhat better results on the whole in those schools which devote most time to the subject. This, of course, would be expected. The important question is: Are the results proportion- ate to the larger time allotments? For example, in the first year the schools of Group III devote 229 minutes more a week to Latin than do the schools of Group I. It appears that the results, how- ever, are only slightly better, not sufficiently superior to justify the larger time allotments. In each of the four years, the schools of Group III devote to Latin about 200 minutes more per week than do those of Group I. It is clear throughout each of these years that the results in the three aspects of Latin ability measured are not sufficiently greater to warrant the increase in time. In other words, it seems probable that the schools of Group III are squan- dering a good deal of time. Results are just about as good with the TIME AND LATIN EFFICIENCY 99 smaller time allotments. This should serve as an indication to the schools of this group that they should reduce the amount of time devoted to the subject and improve their methods of instruction. TABLE 63 TIME ALLOTMENTS AND EFFICIENCY For Pupils Who Have Studied Latin One Year r Number of Minutes Per Week Latin Latin Group School Devoted to Study Grammar Test Sentence Test B of Latin for One Year 17 300 .07 3-41 31 305 .87 3-44 19 350 .50 3-02 Zi 350 1.76 3-50 I 21 375 .60 3.0s 6 390 .28 2.76 3 390 .56 3.23 4 390 1. 16 4.61 34 400 2.24 3.70 Average 361 .73 3.28 26 415 .90 2.35 Z2 415 .50 3.63 29 425 1.72 3-63 16 440 1.20 3.76 n 24 450 1.25 3.24 14 450 .37 3.25 I 500 .30 2.03 23 500 .46 351 10 500 1.09 3.94 Average 455 .96 3.SI 12 500 .87 3-0O 27 520 .16 3,06 9 525 .36 3-20 22 550 1.69 3.38 ni 5 575 2.22, 3-83 11 600 1. 12 3.41 20 600 1.20 4.11 7 640 .91 2.87 13 800 1.37 3.51 Average 590 1.46 3.28 100 LATIN IN SECONDARY SCHOOLS TABLE 64 TIME ALLOTiMENTS AND EFFICIENCY For Pupils Who Have Studied Latin Two Years Number of Minutes Per Week Latin Latin Connected La tin Test Group School Devoted to Grammar Sentence Amount Compre- Study of Test Test B Correct hension Latin for Two Years 2 240 1.27 3-99 19.02 26 17 313 I.I7 3-99 12.99 29 31 338 1.63 3-93 22.83 27 19 375 1.98 3-47 26.69 38 T 15 380 3.01 5-29 58.04 65 6 395 I.16 3-35 12.28 16 3 403 1.49 2.46 13.79 26 21 413 2.17 4-25 31.81 40 i6 440 1.22 4-25 22.11 41 i8 450 •92 3.83 22.14 23 Average 375 1.36 4.04 26.11 36 34 450 2.29 4-23 35-43 54 14 450 .59 3.84 21.99 36 26 453 1.29 3-19 7.67 9 29 463 2.29 5.41 73-40 65 II 4 483 1.64 3.38 18.21 25 25 488 1.85 4.10 34-21 41 24 500 2.39 3-95 23.61 24 I 500 1.66 5.38 32.79 43 23 500 I. OS 3.82 16.86 24 10 500 1-73 6.05 28.04 43 Average 478 1.64 4.01 27.24 34 30 510 1.86 4-15 35.38 48 27 520 .85 484 38.42 68 9 525 1.18 3-57 7.65 12 22 525 2.07 4.69 41.64 53 III 12 525 2.35 4.10 27.14 35 II 570 2.09 4-52 .... .... 5 575 2.79 4.42 60.86 72 20 600 1.81 4-43 19.33 33 7 648 1.09 3-52 27.75 34 13 800 1.72 3-99 38.00 36 Average 580 1-95 4.22 30.54 40 TIME AND LATIN EFFICIENCY lOI TABLE 65 TIME ALLOTMENTS AND EFFICIENCY For Pupils Who Have Studied Latin Three Years Number of Minutes Latin Latin Connected Latin Test Group School Devoted to Study of Latin for Three Years Grammar Test Sentence Test B Amount Correct Compre- hension 2 240 1.92 3-6i 25.38 21 17 359 1.46 4.84 43-99 60 19 384 2.67 4-75 46.60 44 IS 387 2.67 4.27 52.84 92 I 31 389 2.30 4-79 28.33 48 6 394 1.64 4.48 44.51 38 3 407 1.82 3-51 30.22 71 21 425 2.13 4.46 26.70 30 18 450 1.66 4-35 32.20 28 14 450 •23 4-75 29.63 31 Average 389 1.84 4-50 38.09 50 16 452 2.57 5-01 3523 61 26 465 2.28 4.58 23.40 31 34 467 2.59 4.20 56.78 51 24 483 2.52 4.90 57.92 56 II I 500 2.00 5.22 48.12 • 55 23 500 1.72 4-52 43.94 45 29 500 3.01 6.04 44.73 S6 10 500 2.34 4-49 22.12 37 25 515 2.10 4-77 38.11 48 27 520 1-75 5-75 39-09 42 Average 490 2.30 4.80 39-20 46 9 525 2.19 4.04 19.91 28 22 534 2.91 4.81 35.48 46 4 540 2.07 4-35 37.38 39 30 549 2.61 4-38 45.55 59 III 12 550 2.52 4/1 40.56 49 5 575 2.76 5-14 56.70 68 20 600 1.65 4.83 11.38 36 II 605 2.24 4-52 .... 7 639 1.66 4.27 55.28 34 13 800 2.70 5-40 67.02 62 Average 592 2.27 4.58 36.93 41 102 LATIN IN SECONDARY SCHOOLS TABLE 66 TIME ALLOTMENTS AND EFFICIENCY For Pupils Who Have Studied Latin Four Years Number of Minutes Per Week Latin Latin Connected Latin Test Group School Devoted to Grammar Sentence Amount Compre- Study of Test TestB Correct hension Laun for Four Years 2 240 2.06 495 4324 48 19 388 2.76 4.41 49.50 46 IS 390 2.99 5-39 65.12 68 17 393 1. 91 4.86 52.21 56 I 6 393 1.50 4-43 45.62 42 21 4>2 2.03 3.89 42.73 36 31 3 2.26 5-19 34.06 62 3 . 8 1.74 430 23.65 25 i6 4^5 2.07 4.85 33.32 63 i8 450 1.23 4-31 31.38 26 Average 393 1.98 4-74 41.64 45 24 463 2.27 4.66 49.30 46 14 . . . 4.28 44.73 S3 26 471 1.96 4-15 27.63 45 34 475 2.79 S.13 63.20 84 I 500 2.72 5-34 70.89 80 II 23 500 2.18 502 45.34 44 29 519 3.34 5-93 73.51 95 27 520 2.14 4.87 21.58 S3 25 522 2.74 5.14 5439 61 9 525 2.08 499 21.74 48 Average 496 2.40 4.88 45.21 59 22 537 2.7i 4.98 51-57 65 10 538 2.65 524 53.09 67 20 565 3.10 560 38.92 51 30 565 2.74 5-21 58.71 70 III 4 569 2.ii 4.46 37.47 Z3 5 575 3-15 5.80 60.72 70 12 575 2.77 456 4303 43 7 s8o 2.46 S-47 70.86 58 II 623 2.82 5.88 .... .... 13 790 2.66 5-77 8523 66 Average 592 2.75 5-21 5453 56 TIME AND LATIN EFFICIENCY TABLE 67 TIME ALLOTMENTS AND EFFICIENCY Summary GROUP AVERAGES 103 Average Number of Minutes Latin Latin Connec ted Latin Test Croup Per Week Grammar Sentence Amount Compre- Devoted to Test TestB ' Correct hension Latin Study YEAR I I 361I .7i 3.28 .... . . II 455 .96 3-5 1 .... III 590 1.46 3.28 YEAR II .... I 375 1.36 4-04 ; ^ 26.11 36 II 478 1.64 4.01 "^^ 1.95 4.22 ^ 27.24 34 III 580 30.54 40 YEAR III I 389 1.84 4.50 38.09 50 II 490 2.30 4.80 39.20 46 III 592 2.27 4-58 YEAR IV 36.93 41 I 393 1.98 4-74 41.64 45 II 496 2.40 4-88 45.21 59 III 592 2.75 5-21 54-53 56 1 These figures are the averages for one, two, three and four years re- spectivel3^ In the facts brought out here, is another ilkistration of the fact that satisfactory results do not depend so much upon long periods as upon skillful instruction. These schools which get excellent results on relatively small time allotments have adopted methods of instruc- tion which are capable of securing a given result with the least expenditure of time and teaching effort, which constitutes the real essence of economv in education. CHAPTER XIV EVALUATION OF METHOD IN LATIN INSTRUCTION Evaluation of Method Now the important problem in connection with method is to see what the effect of each different kind of procedure in teaching is upon the performance of the pupils. In other words, we need to evaluate the various methods in terms of the ability of the pupils. Does any one type of procedure appear to produce any superiority on the part of the pupils, in ability to grasp the thought of Latin rapidly and intelligentlly ? Schools Chosen for Comparison Out of the entire number of schools, twenty-four were chosen for purposes of comparison of methods. There were some schools in the case of which no comparison would be fair. For example, in the case of a school in which the teaching was poor on account of the inferiority of the teacher it would not be fair to make a com- parison with a school in which the teacher was of superior ability. In the former school the best of methods would give a poor result. Comparisons of method would be fair only in schools in which the same method had been in use during the time that the highest class had been in school. The twenty-four schools which were chosen presented conditions which were similar and comparable to a suffi- cient degree so that the results of such a comparison are valid and fair. In each of the three method groups the schools which were chosen represent the respective methods at their best. In other words, the results secured from these methods in these schools would be expected to be duplicated in any group of schools under normal conditions. In Tables 68-71, the schools are classified according to the methods which they use. It is possible to make three groups on this basis. In Table ^2 is given a summary of the facts in the form of averages of the groups. These figures are a clear indication that EVALUATION OF METHOD 105 in general the translation method is adequate. In each of the three years above the first the schools of Group I give a better result in amount correct in the Connected Latin Test. The superiority of the TABLE 68 COMPARISON OF DIFFERENT METHODS For Ability in the Fundamentals of Latin Class Averages YEAR I Group Method AI AI AI AI I AI BI BII BII Average of all CII CII CII II CII CII CII CII CII Average of all cm cm cm III cm cm cm cm cm Average of all School 6 2 19 i8 15 24 29 30 pupils. 5 13 I 10 3 31 25 16 pupils. 12 9 7 14 4 22 26 27 pupils. Latin Latin Latin Sentence Grammar Vocabulary TestB Test Test 2.76 .28 316 342 •47 3-33 3-02 .50 3-32 ... 3-i6 324 1.25 • • • 3.63 1.72 2.80 •71 . • . 3.13 .75 3-20 3.83 2.33 4.02 3-51 1.37 3.78 2.03 ■30 3.65 3-94 1.09 3.65 323 .56 3-44 .87 . . . 3.96 2.24 . . . 376 1.20 3.65 1-47 3.86 3-00 .87 ... 3-20 .36 2.87 .91 3-20 3-25 •37 3-51 4.61 1.16 3.82 3.38 1.69 2.35 .90 2.99 3.06 .16 . . . 3-13 .88 3-23 schools in Group I over those in Group III is evident in all of the three years in this particular. On the basis of these figures we are justified in asserting that the translation method appears on the io6 LATIN IN SECONDARY SCHOOLS whole to give considerably better results in the schools in which it is in use than does the grammatical method in the schools in which it is used as far as the amount of Latin interpreted correctly is con- TABLE 69 COMPARISON OF DIFFERENT METHODS For Ability in the Fundamentals of Latin Class Averages YEAR II Connected Latin Test Group Method 1 School Amount Amount Compre- Latin Latin Latin Attempted Correct hension Sentence TestB Grammar Test Vocabulary Test AI 6 77.68 12.28 16 3.35 I.16 3.71 AI 2 75.67 19.02 26 3.99 1.27 4.07 AI 19 69.27 26.69 38 3.47 1.98 .... I AI i8 98.69 22.14 23 zs,i .92 3.97 AI 15 88.81 58.04 65 5.29 3.01 .... BI 24 9958 23.61 24 3.95 2.39 .... BII 29 II2.61 73.40 65 5.41 2.29 .... BII 30 74.07 35.38 48 4.15 1.86 .... Average of all '. pupils 89.99 35.13 39 3.95 1.58 3-34 CII 5 84.16 60.86 72 4.42 2.79 CII 13 103.09 38.00 36 3-99 1.72 4.11 CII I 76.87 32.79 43 5.38 1.66 4.52 II CII 10 65.70 28.04 43 6.05 1.73 3.44 CII 3 53.78 13.79 26 2.46 1.49 .... CII 31 84.89 22.83 27 3-93 1.63 .... CII 25 83.29 34.21 41 4.10 1.85 .... Cii 16 54.14 22.11 41 4.25 1.22 .... Average of all pupili 3 72.77 30.88 43.11 423 1.61 3.90 cm 12 77.94 27.14 35 4.10 2.35 .... cm 9 65.15 7.65 12 3-57 1. 18 .... III cm 7 82.53 27.75 34 3.52 1.09 3.89 cm 14 61.02 21.99 36 3.84 .59 3.35 cm 4 73.46 18.21 25 3.38 1.64 3.65 cm 22 78.71 41.64 53 4.69 2.07 .... cm 26 79.30 7.67 9 3-19 1.29 323 cm 27 56.45 38.42 68 4.84 .85 .... Average of ail '. pupil s 72.40 24.16 30 3.81 1.46 348 cerned. Of course, there are exceptions to this statement in the case of individual schools. With reference to comprehension, Group EVALUATION OF METHOD 107 I shows a result superior to Group III in each of the three years. On the other hand, in Group II, the results are slightly higher than those in Group I in two cases. TABLE 70 COMPARISON OF DIFFERENT METHODS For Ability in the Fundamentals of Latin Class Averages YEAR III Connected Latin Test Group Method School Amount Amount Compre- Latin Latin Latin Attempted Correct hension Sentence Grammar Vocabulary TestB Test Test AI 6 115.89 44-51 38 4.48 1.64 4-05 A I 2 119.67 25-38 21 3-61 1.92 4.14 AI 19 105.46 46.60 44 4.75 2.67 .... A I 15 57-67 52.84 92 4-27 2.67 .... I AI 18 116.42 32.20 28 4-35 1.66 4.14 BI 24 103.57 57-92 56 4.90 2.52 .... BII 29 79-33 44.73 56 6.04 301 .... BII 30 77.25 45-55 59 438 2.61 .... Average of all pupils lOI.OI 45.32 51 4.39 2.10 4.12 CII 5 83.00 56.70 51 5.14 2.76 .... CII 13 108.91 67.02 62 540 2.70 .... CII I 85-15 48.12 55 5.22 2.00 4-47 CII ID 60.55 22.12 37 4.49 2.34 425 II CII 3 42.50 30.22 71 3-51 1.82 .... CII 31 58.56 28.33 48 4-79 2.30 .... CII 25 80.22 38.11 48 4.77 2.10 .... CII 16 58.18 35.23 61 5.01 2.57 .... Average of all pupils 74-19 41.42 50 4.80 2.35 4.28 cm 12 82.56 40.56 49 4.71 2.52 • • • • cm 9 72.10 19.91 28 4-04 2.19 .... cm 7 164.95 55-28 34 4.27 1.66 4-49 cm 14 94.96 26.93 31 4.75 .23 4.16 III cm 4 96.60 37.38 39 4-35 2.07 4.20 cm 22 76.72 35.48 46 4.81 2.91 .... cm 26 74.40 23.40 31 4.58 2.28 4.02 cm 27 92.18 39.09 42 5.75 1-75 .... Average of all pupils 93-16 36.21 38, 4.51 2.O4 4.25 In the other aspects of Latin ability the adequacy of the transla- tion method is evident. In Table y2, in connection with ability to io8 L.i77.V IN SECONDARY SCHOOLS translate Latin sentences, there is no case in which the translation method is not practically as adequate as the grammatical method. There is not difference enough between the results of any of the TABLE 71 COMPARISON OF DIFFERENT METHODS For Ability in the Fundamentals of Latin Class Averages YEAR IV Connected Latin Test Group Method School Amount Amount Compre- Latin Latin Latin Attempted Correct hension Sentence Grammar Vocabulary Tests Test Test AI 6 108.92 45.62 42 4-43 1.50 392 AI 2 90.71 43.24 48 4.95 2.06 4.48 AI 19 107.18 49.50 46 4.41 2.76 .... AI i8 120.15 31.38 26 4.31 1.23 3-97 I AI 15 95-94 65.12 68 5.39 2.99 .... BI 24 106.05 49.30 46 4.66 2.27 .... BII 29 77.38 73-51 95 5-93 3.34 .... BII 30 83.88 58.71 70 5.21 2.74 ... * Average of all pupils 103.41 54-12 57 4-72 2.03 4.14 CII 5 87.29 60.72 70 5.80 3.15 4-56 CII 13 129.77 85.23 66 5.77 2.66 4.23 CII I 88.49 70.89 80 5.34 2.72 4.41 CII lO 78.84 53.09 67 5.24 2.65 4.10 II CII 3 95.16 23.65 25 4.30 1.74 .... CII 31 55.18 34.06 62 5.19 2.26 .... CII 25 88.65 54.39 61 5.14 2.74 .... CII i6 52.98 33-32 63 4.85 2.07 .... Average of all pupils 83.45 50.99 61 5.17 2.38 4-35 cm 12 99.71 43.03 43 4.56 2.77 . . . • cm 9 45.22 21.74 48 4-99 2.08 .... III cm 7 121.86 70.86 52 5.47 2.46 4.15 cm 14 84.18 44.73 53 4.28 .... 4.02 cm 4 113.36 37.47 33 4.46 2.33 4.21 cm 22 79.61 51.57 65 4.98 2.73 .... cm 26 61.93 27.63 45 4.15 1.96 4.12 cm 27 40.54 21.58 53 4.87 2.14 .... Average of all pupils 78.87 38.23 47 4.70 2.41 4.12 methods to warrant the statement that one is in any marked degree superior as far as the ability to translate sentences goes. In gram- EVALUATION OF METHOD 109 mar and vocabulary, the translation method gives, on the whole, as good a result, although there are a few cases in which the gram- matical method shows a slightly better average. In these aspects of TABLE 72 COMPARISON OF DIFFERENT METHODS For Ability in the Fundamentals of Latin Summary YEAR I Connected Latin Test Latin Latin Latin Group Method Amount Amount Compre- Sentence Grammar Vocabulary Attempted Correct hension Test B Test Test AI ) I BI BII ) .... .... 3-'i3 •75 3-20 II CII .... ■ ■•• •••• 3.65 1-47 3-86 III cm AI ) YEAR II 3.13 .88 3-23 1 BI BII ) ■ 89.99 35-13 3-95 1.58 3-34 II CII 72.77 30.88 43 4.23 1.61 3-90 III cm AI ) 72.40 24.16 30 YEAR III 3-81 1.46 3-48 I BI f BII ) - lOI.OI 45-32 51 4-39 2.10 4.12 II CII 74-19 41.42 so 4.80 2-35 4.28 III cm AI ) 93.16 36.21 38 YEAR IV 4-51 2.04 4-25 I BI BII ) - 103.41 1 54.12 57 4.72 2.03 4.14 II CII 8345 50.99 61 5-17 2.38 4-35 III cm 78.87 38.23 47 4.70 2.41 4.12 Latin ability, as far as the result itself is concerned, no method has any marked advantage. no LATIN IN SECONDARY SCHOOLS Method in Relation to Time Allotments. The demands for economy of time in education are now so insistent that it is necessary to examine all types of procedure in teaching with strict reference to their time requirements. Of two methods, that which can secure the same or a better result with less TABLE 73 COMPARISON OF DIFFERENT METHODS In Relation to Time Allotments Class Averages YEAR I Number of Minutes Per Week Group Method School Devoted to Study of Latin for One Year AI 6 390 AI 2 240 A I 19 350 A I 18 450 I A I 15 370 BI 24 450 BII 29 42s BII 30 485 Average 395 CII 5 575 CII 13 800 C II I 500 II CII 10 500 CII 3 390 CII 31 305 CII 25 450 CII 16 440 Average 495 cm 12 500 CIII 9 525 cm 7 640 III CIII 14 450 cm 4 390 cm 22 550 cm 26 415 cm 27 520 Average 499 EVALUATION OF METHOD III time, other things being equal, will have to be regarded as better. In view of this fact, it becomes an important part of this investi- gation to examine the different methods in use with reference to TABLE 74 COMPARISON OF DIFFERENT METHODS In Relation to Time Allotments Class Averages YEAR II Number of Minutes Per Week Group Method School Devoted to Study of Latin for Two Years A I 6 395 A I 2 240 A I 19 375 A I 18 450 I A I 15 380 BI 24 500 BII 29 463 BII 30 510 Average 4i4 CII 5 575 CII 13 800 •CII I 500 CII 10 500 II CII 3 403 CII 31 338 CII 25 488 CII 16 440 Average 506 cm 12 525 cm 9 525 cm 7 648 III cm 14 450 cm 4 483 cm 22 525 cm 26 453 cm 27 520 Average 5^6 the amount of time necessary to produce a given result. In Tables 73-76 the time allotments for each school are given and the schools 112 LATIN IN SECONDARY SCHOOLS are classified according to method. A summary of the facts is given in Table yy. TABLE 75 COMPARISON OF DIFFERENT METHODS In Relation to Time Allotments Class Averages YEAR III Number of Minutes Per Week Group Method School Devoted to Study of Latin for Three Years A I 6 394 A I 2 240 A I 19 384 I A I 18 4SO A I 15 387 BI 24 483 BII 29 500 BII 30 549 Average 423 CII 5 575 CII 13 800 CII I 500 II CII 10 500 CII 3 407 CII 31 389 CII 25 515 CII 16 452 Average 5^7 cm 12 550 cm 9 525 cm 7 639 III cm 14 450 cm 4 540 cm 22 534 cm 26 465 cm 27 520 'Average 528 The evidence is unmistakably clear. The eight schools in Group I require an average of 105 minutes, or about one and three-fourths hours less per week for four years than do the eight schools in EVALUATION OF METHOD "3 Group III. These are respectively the translation method and the grammatical method. This figure is secured by taking the time allotments for four years, as given in Table 53. TABLE 76 COMPARISON OF DIFFERENT METHODS In Relation to Time Allotments Class Averages YEAR IV Number of Minutes Per Week Group Method School Devoted to Study of Latin for Four Years A I 6 393 AI 2 240 A I 19 388 I A I i8 450 A I IS 390 BI 24 463 BII 29 519 BII 30 565 Average 426 CII 5 575 CII 13 790 C II I 500 II CII ID 538 CII 3 418 CII 31 413 CII 25 522 CII 16 445 Average 525 cm 12 575 cm 9 525 cm 7 580 III cm 14 469 cm 4 569 cm 22 537 cm 26 471 C III 27 520 Average 531 Of the eight schools which use Method AI, BI, or BII, all but two stand in the lowest half as regards the average amount of time 114 LATIN IN SECONDARY SCHOOLS spent on Latin for four years, as shown in Table 53. Four of these are in the lowest one-fourth, and in fact are the four lowest schools in the whole group of schools. Group TABLE 77 COMPARISON OF DIFFERENT METHODS In Relation to Time Allotments Summary Method AI I BI BII II CII III cm AI BI BII 11 CII III cm AI BI BII 11 CII III cm AI BI BII II CII III cm FOR ONE YEAR FOR TWO YEARS FOR THREE YEARS FOR FOUR YEARS Average Number of Minutes Per Week Devoted to Study of Latin 39500 495-00 499.00 414.00 506.00 516.00 42300 51700 528.88 426.00 525.00 531-00 In the first year, the schools of Group I take 104 minutes less time per week than the schools using the grammatical method. In the second year, the corresponding saving in time is 102 minutes, in the third year, 105 minutes, and in the fourth, 105. There is no great difference in the amount of class time spent in the two meth- EVALUATION OF METHOD 115 ods. In each year, the pupils who are taught by the translation method spend about 100 minutes per week less time studying Latin outside of class than do pupils taught by the grammatical method. Here is a substantial saving of time. A saving of 100 minutes a week for four years is eminently worth while. If the results are equally good by the translation method it will be clear evidence that learning is more economical and efficient by that type of procedure. We have already presented evidence that the results by this method are practically as adequate, and now, when we take into con- sideration the fact that they are secured by conspicuously smaller time allotments, we have a clear justification of this method as com- pared with the grammatical method. We have also further proof of the futility and wastefulness of that plan of teaching which spends so large an amount of time learning paradigms and rules and gives so much attention to syntactical analysis of the Latin texts which are translated in class. One of the interesting and rather significant facts in connection with this aspect of the investigation is the fact that the schools of Group I which devote least time to the study of Latin grammar in a formal way, and, in fact, least time to it in any way, with a con- spicuously smaller time allotment in general, get on the whole as sat- isfactory results in Latin grammar. This is especially significant and is an indication of the source of the great waste which takes place in the teaching of Latin. The Latin Grammar Test was omitted in two of the schools of Group I in the first year and so the results may not be valid. In the second year the schools of Group I get an average score of 1.58 in grammar, while the schools of Group III get only 1.46. Of course there is no great difference between these scores, but the point is that the schools which are putting special emphasis on for- mal grammar, devoting an excessive amount of time to it, and mak- ing it, in a large measure, the end and aim of their work, are get- ting no better results in this respect than the schools which place little or no direct emphasis on this phase of the work. In the third year. Group I averages slightly higher than Group III and in the fourth year Group III is a little higher than Group I, but in both third and fourth years there is no material difference between the ii6 LATIN IN SECONDARY SCHOOLS scores. On the basis of these facts we may say that the grammati- cal method, with all of its stress on formal teaching of grammar, .secures no better result than the translation method with its smaller time allotment in general and its limited stress on the direct teach- ing of grammar. CHAPTER XV ONE OBSTACLE TO SUCCESS IN LATIN AND THE REMEDY Wide Range of Ability in Classes a Hindrance There seems to be no doubt that the fact of the wide range of abiHties found in practically all of the Latin classes at the present time is one of the important factors which contribute to the lack of results. The question immediately arises : Is this condition inevit- able? It does not require a large amount of reflection to see that a very effective remedy has always been at hand, but, as in all other subjects previous to the advent of scientific measurement of results, we have been going on in a blind fashion and incidentally wasting an excessive amount of time and eft'ort. With pupils of all degrees of ability in each of the three upper classes several conditions arise. In the first place, the pupils in the upper quarter of each class are able to translate five or six times as much Latin as those in the lowest quarter. If instruction is adapted to the middle half of the class, those in the lowest quarter can not do the work. As a result they fail for three years. This seems to be a rather long time to take to fail to do a thing. It is certainly a most wasteful procedure. A second result is that those in the upper quarter of the class are allowed to do perhaps only a quarter of what they are capable of doing. Here is an educational waste of the worst kind. Table 78 sets forth graphically the extent of the overlapping in School 38. In Class II there are forty-two pupils, in Class III, twenty-two, and in Class IV, seventeen. All of the pupils in Class III fall within the range of Class II and could well be taught with that class as far as any differences in ability go. Three-fourths of all of the pupils of Class IV fall well down in the division of the scale assigned to Class II. It is evident that if we could select four- teen pupils from Class II, four from Class III and three from Class IV, we could make a new Class II which would be a compact, homo- geneous group as far as ability to translate Latin goes. It would ii8 LATIN IN SECONDARY SCHOOLS >^ H J t—^ m < O Q H ^ C? < l-H tA Q w Psi < u o u o < fH fT| Q en o W (— 1 H o "Z CO 00 H » CO J l-« (X, h4 o Q < w U W o W u CO Id H > O 04 o O U is o OH Uh H H J^ X Q W O C/2 ID < O m CL, CO o 1 — ( 1 — 1 > PL, ^ o ffi ffi u CO ffi o Eh N IN l^ ■* (M i-l 00 iH t^ in i-H l-H l-H tS E " u^ 60 tfl ^ — . ^75 £3 1-. ■»-» 2 -5 'S. 2 « ° - nl -^ g- O u U-- — >*-• 3 J3 w o XI U o ft in nl rt nl (0 .S°t3 °E " S s « ONE OBSTACLE AND THE REMEDY 119 require a rather specious skepticism on the part of anyone to say that these pupils would not do a great deal more effective work in a class together than as they are now. The same would be true of the other two classes. The new arrangement of the three classes is shown in Table 78. There are eighteen pupils who ought either to drop Latin entirely or to go back to first year work. There are four pupils in the senior class who have extraordinary ability as compared with the others. They can probably be assigned additional Latin to be translated out- side of class. All of the others form reasonably homogeneous groups as regards ability and can be profitably taught together. In Table 79 are set forth the same facts for School 41, but the table is based on Latin Sentence Test A. It will be seen in this table that the group of pupils in the second year far overlaps the third year group, and both of these groups entirely overlap the group of fourth year pupils. The conspicuous lack of application of the principles of educational measurement and diagnosis in this school is so obvious that it does not need to be pointed out. A Suggested Remedy For the present we shall leave out of consideration any objec- tions from an administrative point of view and suggest a plan for classification of pupils in Latin above the first year. First, schools in which there are sixty pupils above the begin- ners should have three classes in Latin in addition to those doing first year work. These classes should be made up of pupils of approximately equal ability, on the basis of tests given at the begin- ning of the year, regardless of the length of time they have studied Latin. At the end of the first year of Latin, if a pupil proves to have sufficient ability, he should be placed in the highest group, even. He may remain in this group during the rest of the high school course. Consider for a moment the very much larger amount of Latin which he would translate in these three years in this group as compared with the amount he would get if he remained a year in the second group and a year in the third group. Probably he would translate more than five times as much. Undoubtedly he would cover in this group all of the Latin now taken in preparatory school and the first two years of college. o pa < o o o Q CO LO u w m o H W O in w < > o »— t o LO D O P u < H in Id H w u 2; w H o o u o < 2: o e < CQ 1 H _o n ■(3 o 1/3 '4-* c o V rt u ■5 c c '3 (X, ^4-t O t-i 3 J3 O L.') Od 00 W i-t l-H ^j« f* (C — ' OS 1^ 00 to in o o to I-H o 00 in in o o to iii to IC • i-H rH . • • . W CO in 1— ( t—t >-< i-i p-i . l-H I-. ?► t— 1 <-H .2.-=i;1 ™ -3 < CJ U o 01 .5 « 5 » ^— £? ^ o c w rt fc ra in ONE OBSTACLE AND THE REMEDY 121 Such a plan would necessitate having on hand in each school enough Latin material so that pupils could spend three years in the highest group and not re-read anything. This would not be a difficult thing to do. According to this plan Latin tests will need to be given at least twice a year. Whenever pupils are found in any of the lower groups who have improved sufficiently, they should be placed at once in a higher class, i. e., in a class in which their rate of work will be up to their maximum ability. Only so can the greatest economy be secured. Our present methods in Latin are not succeeding well with more than a quarter of our pupils and the chief reason is the failure to break away from traditional practices and apply the principles of modern scientific supervision and administration. In a good many small high schools, the number of Latin classes is now limited to three by uniting the juniors and seniors and taking Cicero and Virgil in alternate years. In the plan suggested, a given pupil might never get beyond Class II. This would not be a serious matter if his ability never improved beyond that represented by that grade, for the group would have different Latin each year. Con- ceivably Class II and Class IV in a given school might be reading Virgil or Ovid or Livy at the same time, but the difference might be, in the case of Virgil, for example, that Class II would translate six books in a year and Class IV might do the same amount in two months and then pass on to something else. The real basis for dividing the whole group of Latin pupils in the school into divisions would be their ability to do work. It should be stated at this point that this classification of pupils according to ability is sound and valid only on the assumption that their standing in this one test is a perfect criterion of their ability. This assumption, of course, is not strictly true. The writer is fully aware of the fact that the treatment of test results in classifying pupils is a subtle statistical matter which has not been handled adequately to date. But, assuming that by several adequate tests a true and valid measure of the ability of the pupils has been obtained, the idea has a great deal of force in its application to the classifica- tion of pupils in Latin in secondary schools. 122 LATIN IN SECONDARY SCHOOLS Now, are there administrative reasons why this plan can not be adopted? Probably not. If school administration is to be for the sake of reducing educational waste and securing conditions under which learning may take place most favorably, it must find a way by which such a plan may be adopted. CHAPTER XVI COMPARATIVE STANDING OF PUPILS IN COLLEGE Since in many respects the results of the teaching of Latin in the schools included in this investigation have been found to be poor, it might well be argued perhaps that these happened to be schools in which the work in Latin is poorer than in most schools. For this reason, it seemed desirable to determine the standing of the schools in question with as wide a range of other schools, both in and out of the State, as possible. Now, one means of determining the stan- dard of a secondary school is to learn the relative standing of pupils who have gone to college from that school. This was done in con- nection with pupils from twenty of these schools. Of course, they went to many different colleges which have different methods of marking and different standards. The number who took Latin was small and no pupils from some schools took it in college. For these reasons the evidence is far from conclusive, but, on the other hand, it is believed to be of considerable value. From the records collected annually by the State Department of Public Instruction it was possible to determine with entire accuracy how many pupils went to college in the fall of 191 5 directly from the senior class of each of these schools, and the records also showed to what institutions they went. It was found that eighty-two pupils went to college directly from the senior classes of twenty of the schools involved in this test and that of these, thirty-nine took Latin in their freshman year. It was possible to learn the standings of thirty-five of these. The plan was to find the standing in Latin of each pupil from each school for his freshman year and the average standing in Latin of all freshmen in that college. This would en- able us to know whether pupils from these schools were doing as well as students from all other high schools which sent students to these colleges. Table 80 gives the number of pupils from each sec- ondary school who went to college, the colleges to which they went and their average rank in Latin for their freshman year in college. 124 LATIN IX SECONDARY SCHOOLS Number of Pupils Who Number Who Went to Each of the Colleges Indicated Below with Their Average Rank in Latin for the Entered College Directly from pji-gt Year in College. The Figures in the Left Hand Column Under Each College Indicate the Number Senior Classes in Fall of „f Punils Who Attended That College, and the Right Hand Figures Their Average Ranks in Latin. 1915 and Who Took Latin in College J3 o • I- • O • »ft OS • 00 ce 00 in fH . -:-:::::: •* P3 lO CO r4 00 . . • in in in • o 00 o in j o 1 *-* 3 o E u nl Q . CO in u5 ■ • la . . • . t^ la in . ■ «o • • • . . . . o • . . • m . . . . o • • • • o 8 o to . CO 1-1 IN • • (M • • • . . . . r-4 * • • • r-* o tn u V E < in in to m ffj o a in • • • • lO o u o •a o m fH t-H it . • . • in • • • • Oi I' • C • CO • 1- • CO 5 • to • ■* rH * • ■ Mi- ° O M *- OJ am ^ iJ o C Ol-I C J= ■- o « - «2r: 2.-:: '.I u a s c < o n c c ■ n E o o bo < d irt 11 -1 n (0 b£ in en 3 o D. 3 u c K a T3 T3 K V n a> ** V 14-1 D. tn n CO .X u ■*-» l*.l u ri ^ in E (« o C bo n u C a j£ ■" > (11 rt rt > . w u E o OJ 4J n Oi u u >. «J a c h/1 JTl (A 3 OJ HI O 73 o -o f 1 4) D H v u bo se 3^ V "2 n o u u to en o U >> Si aa>^ STANDING OF PUPILS IN COLLEGE 125 It also gives the average rank of all freshmen in Latin. It will be noted that these are colleges which draw students from all parts of the country and for this reason the freshman averages are probably well representative of what the high schools of the country are doing in the development of ability to do college work in Latin. It will be seen that the results are entirely favorable to the schools of our investigation. There are only two cases in which graduates of these schools do not average h%her than do all other freshmen taking Latin. While these data are not extensive enough to be entirely conclusive, yet it is believed that they are a good indication that the schools involved in the test are doing at least as good work in Latin as the average of the schools which are sending students to these colleges. Probably we are justified in concluding that the results which we have found would be likely to be duplicated in the same number of schools chosen at random in any other State. In other words, the results are apparently typical of Latin results everywhere. It is clear that the results, insufiicient as they may be, are highly favorable to the translation method. Schools 2 and 30, which to- gether have ten pupils taking Latin in college, stand very high. School 18, however, draws a low rank for two pupils and School 6 does not stand especially high in the case of one pupil. Schools 26 and 14, both having one pupil, also stand high. School 7 teaches perhaps the most extreme form of the grammatical method of all schools in the list and it will be noted that it stands rather low for three pupils. CHAPTER XVII CHARACTER OF THE PUPILS' ENGLISH A special study was made of the character of the English in what the pupils wrote. This was based on the Preliminary Con- nected Latin Test, so-called, given in 1916. Twenty-one schools participated in this test. To say that the English was poor gives little idea of the disconnected jumble of words which formed the greater part of a very large number of the papers. On the part of fully a quarter of the pupils, there seemed to be no idea of how to read a Latin sentence to get the correct thought and then to express that thought in reasonably good English. A very large number of the papers proved to be entirely meaningless. They were simply words put together with little or no reference to their meaning. It has been claimed that the study of Latin promotes logical thinking, power of exact statement, facility and precision in the use of English and similar abilities. There was no evidence in the papers of these pupils that this is true and an abundance of facts to prove the con- trary, as far as the majority of the pupils were concerned. Of course there were exceptions. Some pupils wrote creditable English. A few schools were reasonably satisfactory in this respect. Generally, however, there was an amount of poor English in most of the pa- pers such that, if the same degree of slovenliness and inaccuracy was allowed in the daily work in those schools, it would be suffi- cient largely to counteract the efforts of the school in all other classes in the direction of good expression. In about a fourth of the schools, as far as the papers of this test are an indication, Latin must be regarded as a study which is about as meaningless to the majority of the pupils as anything can be. What they wrote was to a» large degree incoherent. In another large group, constituting nearly half of the schools, it had little meaning. Preliminary Connected Latin Test The test was chosen, with a few slight adaptations, from Caesar's Civil War. It is reproduced below: CHARACTER OF PUPILS' ENGLISH 127 CURIO'S CAMPAIGN IN AFRICA Isdem temporibus C. Curio in Africam profectus ex Sicilia, etiam ab initio copias P. Atti Vari despiciens, duas legiones ex iv quas acceperat a Caesare, d equites transportabat, biduoque et nocte in navigatione consumpta, appellit ad eum locum qui appellatur Anquillaria, Hie locus abest a Clupea passuum xxii milia, habetque non incommodam sestate stationem, et duobus eminentibus promun- turiis continetur. Curio Marcium Rufum Uticam cum classe prsemittit; ipse eodem cum exercitu proficiscitur biduique iter progressus ad flumen Bagradam pervenit. Ibi C. Caninium Rebilum legatum cum legion- ibus relinquit; ipse cum equitatu antecedit ad Castra exploranda Cornelia, quod is locus peridoneus castris habebatur. Id autem est iugum derectum, eminens in mare, utraque ex parte prseruptum atque asperum, sed tamen paulo leniore fastigio ab ea parte quae ad Uticam vergit ; abest derecto itinere ab Utica paulo amplius passus mille. Hoc explorato loco Curio castra Vari conspicit muro oppidoque coniuncta ad portam quae appellatur Bellica, admodum munita natura loci, una ex parte ipso oppido Utica, altera theatro quod est ante oppidum. His rebus gestis Curio se in castra ad flumen Bagradam recipit atque universi exercitus conclamatione imperator appellatur, pos- teroque die Uticam exercitum ducit et prope oppidum castra ponit. Nondum opere castrorum perfecto equites ex statione nuntiant magna auxilia equitum peditumque ab rege missa Uticam venire; eodemque tempore vis magna pulveris cernebatur, et vestigio tem- poris primum agmen erat in conspectu. Novitate rei Curio permo- tus prsemittit equites qui primum impetum sustineant ac morentur; ipse celeriter ab opere deductis legionibus aciem instruit. Equitesque committunt prcelium, et, priusquam plane legiones explicari et con- sistere possent, tota auxilia regis impedita ac perturbata, quod nullo ordine et sine timore iter fecerant, in fugam coniciunt; equit- atuque omni fere incolumi, quod se per litora celeriter in oppidum recipit, magnum peditum numerum interficiunt. 128 LATIN IX SECONDARY SCHOOLS The twenty-one schools were arranged in three groups as regards the adequacy of the expression of the thought of the Latin in English. The following is the result : Schools in Which the Expression of the Thought was Very Inadequate as Regards English Schools in Which the Expression of the Thought was Poor as Regards English Schools in Which the Expression was Good as Regards English 9 21 31 2l 26 22 i8 I 14 5 24 19 25 23 3 U 22 7 13 IS 6 36 1 Year II only. 2 Years III and IV only. School 2 fell in two groups. The second year class, which had been taught by the direct method, was conspicuously poor in English, while the other classes did very well. It is interesting to note that no school in which the pupils were taught by the translation method, falls in the group of schools in which the English expression is poorest. All of the schools in the group in which the English is most inadequate, with the exception of one class of School 2, teach the grammatical method. Of the five schools which teach the translation method, three are in the group in which the English is best and two are in the middle group. Below are given in italics correct translations of a number of sections of the Latin of the test followed by characteristic English statements of the same thought found in the pupils' papers. These are simply a few examples chosen at random to illustrate the kind of inaccuracy found all through the papers of more than two-thirds of the schools. CHARACTER OF PUPILS' ENGLISH 129 At the same time Curio set out from Sicily into Africa. At the same time C. Curio waging war in Africa from Sicily also. For the same time C. Curio in Africa sent from Sicily and now by waging war. At the time C. Curio in Africa having gone from Sicily. At the same time C. Curo, a profectur, in Africa from Sicily. This — ^C. Curo in Africa from Sicilia to . It has in summer a convenient harbor. It is hemmed in by two projecting promontories. He had not spent the summer at this station and he continued by two eminent promotions. This place was not near a station, and by two eminent personages was continued. They took up their station for summer; it contained two enemys. The inhabitants had no real station. He had no means of a guard in summer, and two being emitted he con- tinued with the rest. It was held by two eminent promoters. It had not an insufficient summer station and was held by two generals. He had not stationed commodations by summer, and by two eminent pro- muntaties he filled. It is held by two prosperous tradesmen. After two days and a night he landed at Anquilaria. The cavalry was transported at Bidue and stayed in ships. And the twelve night was taken up in navigation. And by the middle of the night spent in navigation he called to this place which is called Anquillaria. He having consumed the ships during the night he called to this place which they called A. And twice by night in navigation went, he called to them which were called Anquillaria. This zvas on account of his scorn of Icarus' troops. Also Publius Various from friendly aid. And now by wageing war P. Atti Vari despised. Also dispatching P. Attius Various. The land P. Atti Vari describes. Then he scattered the troops of Publius Attus Varrus. From the beginning the forces of P. Attus Various dispersing. Also Publius Atti Varius by beginning to desipline troops. P. Atti Vari selecting troops from the forces. And to the initial troops of P. Attuis Varius — 130 LATIN IN SECONDARY SCHOOLS It is a straight ridge extending into the sea. This however is a direct alliance, prominent in the sea, and fierce and barbarous in this part. This however is eminent in ships, and also praeruptum from the part and aspered, but wearied by so great a leniore he hastened from this part to Uticam. This also is directed, enemies on sea, and from a part waves and tide, but however having got a little way. This however is East in direction prominent in the sea farther from this eruptive and wild part but nevertheless he approached slowly. This thing however was given judgement, near to the seas and but never- theless fatigue was small from those who had turned to Uticam. to This yoke however he is directed going over the sea in all directions rough and storms but however. This, however, is directly just, eminent in the sea and towers and aspires. The other one directed the yokes, with the same slow steps he turned toward Uticam. However in the direction of the march, embarking in a swamp and pushed on from other parts but nevertheless a little worn out. Ctirio sent Rufus to Utica with the fleet. He sent with this fleet Curio Marcium Rufum Uticam. Curio Marciam Rufum Uticam with the class before stated — Curuus IVIarcius Rufus was sent ahead to Utica with the class. Curio Marcuum belonged to the Rufum Uticam class. The following are typical samples of the incoherent English found all through more than half of the papers, and show the kind of force, precision and exactness of statement, or rather the lack of these characteristics, which the study of Latin develops. The same time C. Curio in Africa, went from Sicilia then from Sady P. Atti Vari, — two legions from IV — accepted by Caesar horses transported, — by night — to this place who was called Anquillaria. Curio Marcium Rufum Uticam. with . There C. Canuim Rebilum stayed with his army. But it was in the sea. But however from this part shortly. The same time C. Curio from Sicilia from Africa two legions. Curio Marcium Rufum Atticum when it was permitted to class himself with the same army progressed along. At the same time C. Curio in Africa set out from Sicily, not by initiating the coupplies of P. Attit Vari to—, two legions from IV was accepted by Caesar, D the cavalry was transported, and by night and day, he called thio place after Anquillaria. CHARACTER OF PUPILS' ENGLISH 131 Samples of Pupils' Papers Several samples of the pupils' work, consisting of complete papers, are reproduced on the following pages. Papers were chosen which are representative (i) of the best, (2) of those which are good and (3) of the poorest. The Best Papers Here are five which were among the best and which were very rare in any school. In fact, only a very few such papers were found in the entire group. A. At the same time Carus Carius set out from Sicily into Africa, des- pising wen from the beginning the troops of P. Attus Varus. He took with him two of Caesar's legions and 500 cavalry and he arrived at a place called Anquillaria in two days and a night. This place is 22 miles away from Clupeam and it has a convenient station for summer and it is bounded by two large promontories. Curio sent Marcius Rufus Utica ahead with the fleet; he himself took the army and arrived at the river Bragada in two days. There he left the lieutenant Caius Caninius Rebilus with the legions ; he went ahead with the cavalry to explore the Cornelian camp, because he liked this place as a site for a camp. However it is on a ridge, rising up in the sea, on both sides steep and rough, but sloping a little more gently towards Uticum ; it is a little more than a mile distant from Utica in a bee line. Curio explored this place and saw the camp of Varus which was joined by a wall and the town to a gate which was called Bellica, a place very well fortified. B. At the same time Cassius Carlo setting out into Africa from Sicily even seeing the forces of Publius Attus Varus from the beginning called to this place which is called Anquillaria, two legions and the four which he had expected from Caesar, he led across five hundred cavalry and used up day and night in navigation. This place was distant from Clupea twenty-two miles and it did not have an unconvenient station in summer, and it was bounded by two eminent promentories. Curio sent ahead Marcius Rufus Utica with the fleet; he himself set out with the same army, and advancing o two day journey came to the river Bagradam. Here he left C. Canimuin Rebelus a lieutenant with the legions ; he himself went ahead with the cavalry to explore Castra Cornela because this is held in by a camp. This, however, was a straight ridge, near the sea and on both parts sharp and asper, but nevertheless a little gentler sloping from this part which lie toward Utica; it was distant from Utica in a direct journey a little mor than a mile. This place being explored. Curio saw the camp of Vara, joining the wall and town to the gate which is called Bellica, fortified in this manner by the 132 LATIN IN SECONDARY SCHOOLS nature of the place, on one side by the town of Utica itself, by the other, a theater which is before the town. After these things were done, Curio retreated to the river Bagrada and also the whole army called him general by a shout, and on the next day, he led his army to Utica and pitched camp before the town. The work not yet being completed, the cavalry announced from their post that a great aid of cavalry and footmen sent by the kind came to Utica. At the same time great strength of was seen, and the flag of the first line of battle were seen. Curio moved by this new thing sent ahed the cavalry •who should sustain the first attach and delay; he himself quickly leading out the legion from the work drew up the battle line. The cavalry began a battle and before the legions could be plainly seen and take a stand, all the aid ■was hindered and disturbed, because they made the journey without any order or fear. They C. At the same time, C. Curio set out into Africa from Sicily, likewise from the beginning sending the forces of P. Atti Vari, he had received these two legions from Caesar from four; he had transported five hundred horse- men, for two days and a night they were on the sea, he called those to him who were called the Anquillaria. This place was distant from Clupea twenty- two miles not being used as a station in summer and covered two promon- tories. He put Marcium Rufus Utican in charge with a fieet ; he himself set out "with an army and, marching for two days come to the river Bagrada. There he left C. Cananium Rebilum a lieutenant, with the legions. He himself pro- ceed to camp because this is the place where the D. At the same time C. Ceerius having set out from Sicily into Africa also the troops of P. Attuis Varius from the beginning, he brought across 2 legions from (horsemen f) this four which he had received from Caesar, and lOO horsemen, having taken 2 days and a night in sailing, to this place which was called Anquillaria. This place was 22 miles away from Clupea and it has not an inconvenient situation f at summer and is held in by two emi- nent promuntories. Curius sends ahead Marcus Rufus Uticus with the fleet. He at the same time sets out with the cavalry and hav^ing advanced for a two day journey he came to the river Bagradam. He left there C. Caninuis Rebilus, a leeu- tenant with these legions. He himself proceeded with the cavalry to Castra for exploring Cornelia, because this place is held by a camp. E. At the same time when Curius set out from Sicily to Africa, P. Attius Vari also received from the first forces two legions from Caesae, he carried a cross five hundred cavalrj'men and the night being consumed in sailing he came to the place which was called Anquallaria. This place was about twenty- two miles from Clupea and it lead the commodities of a winter station and it held the promentaris of two emuuent positions. Marcius Rufus ivas placed in charge of the fleet by Varius, he himself, set out for with the army and came to the iOd which leads to the Bagrada CHARACTER OF PUPILS' ENGLISH 133 river. There he left Carius Rebilus the lieutenant with the legions; he him- self went ahead with the cavalry to Gastra which had been explored by Cornelias which place for a camp they had. This, however, is a direct alli- ance, prominent in the sea and fierce and barbarius in this part, but how- ever, he Papers Representative of a Majority The following papers are representative of the work of a very large number of the pupils and are typical of a majority. They were found in all years. A. At the same time C. Curio a profectus in Africa from Sicily and from the army of P. Atti Vari. Received two legions from Caesae, howseman were transported, and at night the voage was completed. The Anquillaria were called together at a certain place. This place was about 12 miles from Clupea and had a good summer quarters and here they were held. Curio Marciaim Rufum Uticam with the class before stated was at this time with an army making a march to the Bagradam. Here C. Caminium Rebilum a legate remained with the legions. The same was sent with the horsemen to the camp of Cornaliua. he however was ordered to cross the sea to Utican and to march from there to a part a few miles from Uticam. The scouts sent ahead saw the walls and the town which B. The same time Casium Carius had set out from Africa into Sicilia with two legions from the forth which had come to Caesae, the soldiers were transported, and in two days and two nights had departed in shyis, he went with them to a place which was called Anquillaria. This place was twenty thousand miles from Clupea. Caeius Marcius Rufum, was sent with a fleet to Utica, he himself with the same army had set out and two day after he crossed the river Bagradam. When Caeius Canninuam Rebilum ambassador was left with the legions, he returned with the soldiers into camp, because this camp was in a dangerous place. He however set it on fire, and C. In a short time, C. Curio Africa departed from Sicilia accompanied by some troops. P. Atti Vari, two legions accepted by Caesae from the forth. P transported the cavalry the second decided to navigate in the night. He was called to a place which was named Anquillaria this placabest by Clupia passed twenty-two miles, having made the station in summer and two held eminintitus promuntarus. Curio Marcus Rufus Uticam D. At this time C. Curio had come into Africa out of Sicilia, likewise he took the forces of P. Atti Vari, two legions out of the four which Caesar had he also brough 1,000 horses and by night he entered the place in a ship he stopped at a place called Anquillaria. This place is 2,200 miles distant from Clupea and he had to station his troops for the summer. 134 LATIN IN SECONDARY SCHOOLS Curio sent Marcuim Rufum. Uticam forth with troops also with an army and on his march he came to the river. E. At this time C. Curius set out into Africa from Sicily. He had the forces of P. Atti Varus to legions from the four which he had expected from Caesar and 500 cavalry to transport. After two days and a night of siling, he called at that place called Anquil- laria. This place was 22 miles from Culpea, was not inconvenient for a sum- mer station and was surrounded by two overhanging promontories. Curio Marcius Rufus Uticam had come before him with the fleet. The Poorest Papers The following are a few papers which represent those which are poorest, and which were found in large numbers in all years. A. This — C. Curo n Africa — from Sicilia, to to legions from Caesae were taken, the equites were transported by night in the ship, they were called to this location, which was named Anqullaria. This was about twenty-two miles. Curo Marcuim Rufum Utum when he command his army to progress to the river Bagradam set out before them. There C. Canium Rebilum the lieutenant he met wath legions, — then came to the camp of Comeila which was in this location. B. At this time C. Curio went from Sicilly into Africa. He took two legions from Caesar four the horses were transported, and started at night to a place which was called Anquillaria. This place was twenty-two miles from Clupea Curio Marcium Rufum Uticam with C. At the same time Ciauus Curius set out from Sicili into Africa also from the first foras Pilius Atti Vari, with the two legions from the four which he had received from Cauar called him to a place which was called Clupea by twenty two miles and D. At the same time Curuis Curio waging war in Africa from Sicily also. He was accepted by Caesae with two legions he transported the horse- men and having consumed the ship during the night he called to this place which they called Anquillaria. This place wos twenty-two miles from Clupea. Continued by two eminent promontory. He sent with this fleet Curio Marcium Rufum Uticam. He himself when he had fitted out the E. At the time of C. Curio in Africa having gone from Sicilia with a few troops with P. Atti with two legions out of five accepted by Caesar, the cavalry was transported at and stayed in ships. The name of the place where they stayed was called Anquillaria. This place was clear for twenty miles. The inhabitant had no real station it two When C. Caninium rebelian ambassador with legioner reaming who when the cavalry. CHARACTER OF PUPILS' ENGLISH 135 F. At this time C. Cruio had come into Africa from Sicily, two legions from four which was accepted by Caesar, one hundred horses he had trans- ported. This place is by Clupea 22, — miles, and had not stayed there in summer, and two Curio Marcus Rufum Uticam he sent before with a class; they them- selves It should be remembered that each of the above is a complete paper. Of course superintendents of schools and principals of secondary schools are fully aware of this state of things in Latin classes. If practice in writing good English is essential to the habit of good English expression, Latin as taught is a positive detriment to the learning of English. CHAPTER XVIII THE TEACHING OF LATIN General Point of View It is the purpose in this chapter to present a summary and inter- pretation of the facts brought to Hght in the foregoing seventeen chapters. It is not the plan, as the title of the chapter might imply, to enter into a pedagogical discussion of the subject in the sense of offering an outline of a specific method of teaching Latin or even suggestions along that line, except to the extent that they are in- volved in a consideration, in summary form, of the body of signifi- cant information, as a whole, disclosed by this investigation. It is felt that the study ought not to be brought to a close without an interpretation of the results in the light of all of the facts, although at various points, as the discussion has advanced, the outcome of different lines of inquiry has been considered somewhat incidentally in connection with the data at hand up to that point. It has not been a part of the plan of this study to consider the question of the educational values of Latin or the justifiability of teaching it or its place in the curriculum or the proper time to begin it or problems of that sort. It has been necessary to limit the inquiry to the specific matters which are discussed in the preceding chapters and which center around the problem of the present effi- ciency of the work of the secondary schools of one State, in this subject. The following statements are recognized as tentative conclu- sions. The facts brought to light in this study indicate that they are true, but of course they are subject to verification by other in- vestigators. Specific Interpretation of the Results as a Whole I. The teaching of Latin at the present time is failing with a majority of the pupils in secondary schools. Anyone who, after reading the papers written by the pupils concerned in this investi- THE TEACHING Of LATIN 137 gation, would be disposed to claim that high schools in four years of instruction succeed in developing a real and permanent mastery of Latin would be, indeed, pursuing a fantasy. The evi- dence is conclusive that they are not doing so. On the other hand, there is no proof that a mastery of Latin to the extent of the acqui- sition of an adequate reading knowledge of the language and an efficient knowledge of the simple and common aspects of grammar is incapable of being attained in four years. It must be said, however, that in a number of these schools Latin is taught under about as favorable conditions as could well be found and some of them secure with some pupils reasonably satisfactory results, but the majority do not succeed in developing a degree of efficiency on the part of a sufficiently large number to justify the time and effort devoted to the subject. After having found the outcome so unsatisfactory it is impor- tant to point out the exact source of the failure. To fail thus to interpret the results of the study would be to leave it incomplete. In the following pages, an attempt has been made to do this. The inadequacy of the work in Latin comes from three princi- pal causes: (i) lack of adaptation of the subject as now taught to the needs of adolescent youth; (2) absence of the application of well-recognized principles of administration and pedagogy; (3) poor choice of the Latin material constituting the content of the course. 2. Great waste occurs in Latin instruction through the ineffi- ciency and wrong emphasis of present methods of teaching. De- tailed analysis of Latin classics is carried on with an intensity and thoroughness which are beyond all reason. Latin classes are more like clinics than courses in which inspiration is gained from study- ing portions of a great literature, selected with reference to the needs, interests and powers of appreciation and understanding of boys and girls. There is an extreme of pedantic persistence in re- quiring a mastery of unimportant details of language construction in which the pupils never really become adept. Fine distinctions and difficult abstractions, of the nature of logic, which are beyond any genuine comprehension or competent mastery by adolescent boys and girls, are insisted upon even to the point of producing eflfects which are intellectually deadening. Pupils are held rigidly 138 LATIN IN SECONDARY SCHOOLS to account for syntax in which these things are the chief goals, often at the expense of many other matters of a distinctly human interest which youths are able to understand and from which they might derive real educational value. There is vastly too much insistence on drill and routine and far too little on inspiration. 3. This kind of instruction in Latin is of a character which is especially poorly adapted to the needs of the pupils. According to our present conception of the intellectual needs of the period, adol- escent youths are least fitted of those of all ages from babyhood to ripe manhood for the kind of detailed and abstruse linguistic analy- sis previously mentioned, the emphasis upon which makes Latin a formal subject lacking in a content which is in any marked degree significant to the pupil. The structure of the language receives the chief emphasis at this most inappropriate period in the pupil's life. Thus, these aspects of Latin are barren for the pupils, unproductive of any real results as instruments of growth in their lives, useless from an educational point of view, and the attention to them is thoroughly unjustified by modern educational science. Spirit and content are sacrificed for structure and system to an extent which is pathetic. 4. Aims in teaching Latin are vague and in many cases unjusti- fiable. It is not the purpose to enter into an extended discussion of this problem. Certain things, however, are obvious. In about a third of the schools concerned in our test the evident aim of the work is to develop ability in syntactical analysis of a Latin text. It is clear that this is the purpose, for a very large pro- portion of the total time given to Latin is devoted to it. The evi- dence clearly shows that this gives no superiority in functional knowledge of the constructions of the language. If the purpose, or at least one of the main aims of the teaching of Latin, is to develop the ability to read or translate it eflfectively, the end in question is not a valid one. Many teachers will claim that the chief value of Latin comes, not from an ability to apprehend eflfectively the thought of the lan- guage, but from the type of mental training afiforded by this lin- guistic analysis. Suffice it to say that modern thought in education does not justify such a view to the extent to which it is often held. THE TEACHING OF LATIN 139 5. The usual procedure in teaching in the first year is uneco- nomical. The facts brought out in the previous chapters suggest rather emphatically that beginning the study of a language by first mastering a text similar to the ordinary beginners' book in Latin is highly wasteful. According to the plan of most books of this kind the various les- sons present grammatical principles which are explained and illus- trated by sentences chosen for this purpose. In connection with this are lists of Latin words with their meanings, to be memorized. Following these is an exercise consisting of Latin sentences to be translated and English sentences to be written in Latin. These usually illustrate, mainly, the principles of the day's lesson with some review of previous work. There is a limited amount of con- nected Latin material for translation scattered at intervals through- out the book and a number of pages of stories at the end. More than four-fifths of the book, however, consists of the systematic presentation of the principles of syntax together with the illus- trative sentences and the English and Latin exercises. The emphasis is entirely misplaced from the point of view of economical learning. To begin the learning of anything by first mastering the principles upon which it is based, is a form of pro- cedure which is being looked upon with decided disfavor by those who have made the most careful study of children's learning. It is quite clearly recognized by students of education that some facility in dealing with any particular subject purely as an art should be acquired at the outset before any extended study is made of the principles on which it is founded. This would require that pupils learn to translate Latin somewhat effectively, — just as they learn to talk or to read English, — before entering upon a detailed and more or less abstract study of the grammar of the language. Time and public funds are now wasted in an extravagant manner by those schools which still cling to the traditional method of teaching first year Latin. Sufficient evidence has now been accumulated, based upon experience in the class-room, to furnish conclusive proof that, after a few introductory lessons, pupils can immediately begin to trans- late simple Latin sentences, that they may continue the translation of easy Latin after once getting started, and acquire familiarity 140 LATIN IN SECONDARY SCHOOLS with new forms and principles of syntax by meeting them again and again in the situations in which they normally function, and that they do not need to be able to describe constructions, classify forms and state, in a formal way, principles of grammar in order to be able to react correctly to them in sentences. There should be, then, an abundance of easy and interesting Latin material at hand from the beginning and pupils should trans- late a very much larger amount during the first year than is now the custom in most schools. In fact, the main emphasis should be put on this side of the work rather than on the formal study of gram- mar. Learning by doing should be the predominant aspect of method. Typical procedure in teaching in the best schools represented in this study corresponded closely to the following. During the first year the equivalent of the amount of Latin found in several ordinary beginners' books was translated by the class. This was done with no preliminary study of grammar. The school placed at the dis- posal of the teacher a number of sets of beginners' books, — chosen largely with reference to the amount of connected material in them, — and also several sets of easy Latin story books. The teacher followed no particular book but exercised her function as a teacher in presenting orally to the class, with the use of the blackboard on which to illustrate, the minimum knowledge about usage which the pupils required, when it was needed. Probably considerably less than ten minutes a day was necessary for this and the rest of the time was spent, from the beginning, translating connected Latin. The teaching consisted almost wholly in explaining and illustrating to the class facts of usage. Pupils were never burdened with tech- nicalities and the entire stress was put upon functional learning. It is believed that one of the first steps necessary in a reform of Latin teaching in secondary schools is to break away from the for- mal and traditional methods which have prevailed in the first year in the past. 6. The use of the class period in the tipper years is uneconomi- cal and ineffective. A few schools translate all of their Latin at sight, using the class period practically exclusively for this purpose, and several others require a good deal of sight work. By far too many teachers, however, waste this time in listening to slow. THE TEACHING OF LATIN 141 labored, word-by-vvord, literal translations which the pupils have with difficulty worked out by themselves outside of class, to recita- tions of rules of syntax in connection with the constructions of the day's lesson and to descriptions and classifications of Latin forms. All too rarely does one find the teacher's conception of the class- room to be that of a workshop in which teacher and pupils are working together, teaching and learning how to study effectively. A large part of the failure in Latin is doubtless due to this lack of a proper conception of his work on the part of the teacher and also by those administratively in charge of the schools. Just as pupils are taught how to study in connection with read- ing in the grades, so there should be definite training in the high school Latin classes in getting the thought rapidly and efficiently. There should be a great deal of practice in glancing rapidly over a sentence or a number of sentences or even a paragraph, sensing the meaning in the original, followed by a statement of the thought in English. Pupils would thus establish habits of grasping accurately the thought as they do in English, without being focally conscious of all the minute particulars of form and syntax. Facility in dong this well would enable Latin classes to read pages where they now read sentences and whole books where at present they read chapters. This is one of the important causes of the meager, results in teaching Latin, 7. The ordinary methods of study on the part of pupils are very uneconomical and unsatisfactory from a pedagogical point of view. Time and energy are needlessly and lavishly squandered in the use- less and unprofitable hours spent by pupils outside of school "thumbing" a Latin dictionary in "getting" lessons. Lesson-setting and reciting predominate to the exclusion of real teaching how to do by doing, in the class-room. In not more than three or four of the schools, is any effective effort made to teach, in class, how to translate. The method of simply assigning a lesson and leaving pupils to get it by themselves predominates to vastly too great a degree. The pupils' written papers in this test show the results of such unscientific procedure in their lack of ability to attack a simple Latin passage to get the thought of it effectively and put it into coherent English. Detailed notes in the text books, dealing largely with form and even philology, exist in profusion and are included 142 LATIN IX SECONDARY SCHOOLS in the assignment of the lesson, often to be studied in the evening at home, amid all sorts of distractions. Pupils are required to "prepare their syntax" with precision and exactness in their care for the most minute details of form. Rules must be learned and constructions classified and, after being kept under the most thor- ough instruction and rigid drill for four years, a little over fifty per cent accuracy in knowledge of construction is the best that pupils can do. 8. The present teaching of prose composition contributes almost nothing to ability to read or translate Latin. The evidence indicates that here is an important source of waste. Secondary school pupils never reach such a point that they can write Latin automatically. The pupil is obliged to hold in mind, as he writes, and consciously apply a multiplicity of facts of form and syntax, and as a result his writing in Latin is slow, labored and difficult. In composition, one does the reverse of what he must do in order to read well. He who is obliged in reading to give attention to textual details can never interpret the thought well and rapidly. There is plenty of proof that it is possible to read a language effectively without being able to use it in writing. The evidence growing out of this investigation throws serious doubt on the value of Latin composition. May it not be that in the early stages of learning Latin, — say the first three years, — it is a direct hindrance to the process of acquirng a ready and effective grasp of the thought of the Latin? Certainly it appears to do little or no good as far as its main alleged purpose is con- cerned, i. e., to clarify principles of syntax in order to enable pupils to translate better. The plain fact is that three years of teaching and drill and a large time expenditure on the part of the schools which emphasize it most, produce no particular superiority in ability to translate or in knowledge of construction. 9. It is difficult to understand hozv the present teaching of Latin can have anything but a bad effect on pupiJs' English expression. In making this statement we are judging wholly on the basis of the kind of English which they used in their papers. The majority of the papers were seriously defective from the point of view of English composition. If these papers represent the pupils' habitual method of use of English in translation, the study of Latin is about THE TEACHING OF LATIN 143 as well calculated as anything could be to prevent the development of any habits of vigor, clearness and correctness of expression. Observation of the work in Latin classes leads the writer to believe that these papers do represent a use of English in transla- tion which is nearly universal. Such a thing as good sentence- sense on the part of pupils and excellent, clear-cut sentence structure in translation, as far as these papers are concerned, is extremely rare. 10. A large part of the present failure in teaching Latin is due to the lack of application of principles ivell known in elementary school practice zvhich are nozu finding wide application. Almost nowhere in the schools concerned in this study are Latin teacher? effectively diagnosing class and individual needs. Great educational waste occurs at this point. Many pupils of superior ability work for several years in classes in which the majority of the members are far below their ability and thus they do only a half or a third of what they are capable of doing. On the other hand, pupils of in- ferior ability are kept in classes in which the majority of the pupils are four or five times as capable as they are. Thus, they fail con- tinually. All of the Latin classes above the first year represent all grades of ability. By properly classifying the pupils into divisions on this basis, the working ability of the entire group of pupils studying Latin doubtless would be increased several fold. An edu- cational system which pretends to be founded on scientific principles can no longer tolerate such a wasteful practice as now prevails in this particular. 11. There shotdd be a better choice and more interest and variety in the Latin material. The traditional round of Caesar, Cicero and Vergil, with a year devoted to each, is particuluarly un- suitable material for the secondary course. In the first year, there should be a good deal of easy material for translation or reading, such as stories, fables, history and mythology. In the other years, a wide range of authors should be covered. Some of the schools among the number tested study, above the first year, Cicero's De Senectute and De Amicitia, Eutropius, Tacitus' Germania and Agricola, Nepos, Horace, Ovid, Pliny and a great deal of similar material. One school devotes nearly all of its senior year to authors usually studied in college. 144 LATIX IX SECONDARY SCHOOLS Our present method of teaching Latin, as far as the content of the course is concerned, would be hke attempting to teach EngHsh to thirteen or fourteen-year-old foreigners by a purely grammatical method based upon study of abstract linguistic principles. At the outset there would be a year of severe and unremitting drill on the principles of grammar, in a highly abstract form, with the facts illus- trated by formal exercises, consisting largely of isolated sentences, supplemented by a limited amount of reading of connected material. Following this would be three years more of study, almost entirely from the point of view of language constructions, philology, prin- ciples of prosody and such matters, of very limited portions, — thirty or forty pages to a year, — of Grant's Memoirs, Webster's Orations and Milton's Paradise Lost. This would be accompanied by still more study of grammar, combined with the writing of many iso- lated sentences and a few connected passages, for the purpose of clarifying the construction of each writer. Such a procedure is only a little short of pathetic. No one would expect boys and girls to gain a useful command of English reading or the appreciation of literature in that way, nor would it be probable that they would even master English grammar by such a method, no matter to what extent it was made the end and aim of the course. Conclusion By no means are all of the defects discussed in this chapter common to all schools, but they are sufficiently widespread, in the judgment of the writer, after several years of study of Latin in- struction combined with the facts brought out in this investigation, to be the cause of the present unsatisfactory results. There is no school in which none of these inadequacies can be found, and they exist in varying degrees in all of them. If Latin is a subject which is capable of mastery by high school pupils, this end could probably be achieved by any school, under normal conditions, which succeeding in freeing itself from all of these faults, and it is quite likely that it would be necessary for a school to rid itself completely of them before it could expect to at- tain an entirely satisfactory result with an economical expenditure of time and effort. APPENDIX A 145 n H H tn O H w O » u U tij 1^ 1< H t« r/1 n H ^ « en 1— t F-. o ^ H t/l c 1 < H w u CO 05 O < o o '/J :^ s ^i] < •A h3 p^ H 'A ^ 5 H < H CJ «-^ H w > ^ rt in H H H I — I < -♦-J en 5 tn lU u C i r^ ' -rHrHC^rHMO 'rHi-H •■<»» Ci COr-( *rH05 'i-H-* *-*'Mr-tMCOt^O^CO 'r-ICOi— 1-^ 'OS • • -W^ii^ O 00 O O O I/! C"]'^r-] .-1 1— ' !>. rH rH CO 1-f rH (M ■ i-H ■ r-. I* • (>J l^ . »-t .-H OOr-t -i-H •-*i-( * '^ • • - • rHr-l)-tr-»r- — r- — 1— t'Mi^l(M<>J w rt H (« Z w Ui CO l- :2 CQ 1) »— ) H u C 1 — 1 u (/) >— * fTl u H 25 •*- Pi < S ^ < , 1 2; I- w t/) '^ c < ^ ^ « H ^ H < t^-i H >H 03 H H •— 1 h4 n t— t CQ ■*-* < 3 o o 01 . . . Tf • fM • i-H C^ 1 (M ■ • (M r- '•-J^i-iCO •i-ti-i'MT— iri •i-'F-ICC' t-HiH • -ClrH •(MrHrHCO-* 'iTl M< i-t • r' ■M rH • 00 iH iH >XCO(MG^c-lCO-*lO tl O u •a c rt HH H CJ < (A! W H c W w CO < A u r/0 J2 '^ F < Z s Cii H :^ < >H H H l-H J n (— * n <: tn u C C u IN 00 CO 1— < • IN 0-1 rH r-t rH r-t CO rH r-t r-i r^ CO r-t rH r^ rH CO rH 1-t 1-* CO r-t rH r-t a rH rH rH r^ r^ r-t r-t r-t rH r-t o> o» rH rH r-i r^ eJ in rH rH IN r-t rH rH rH 00 rH r^ 1^ r^ r~{ «5 rH . e^ r-t rH CO (N rH IN t~ rH . IN r-. r-t r-ie-M^ <0 ■* IN rH CO eq rH CO IN in IN ■^ ■^ !0 rH r-H • rH rH rH • w^r^ j-t 10 TKeq CO IN IN Hji rH rH IN CO CO rH t- ■* in rH W ^ rH rH • (N rH rH (N Cq CO r^ CO * CO IN IN IN « rH t~ m r-t (>i 1-t IN CO IN o> e-i '^ f^i r-^ rH C-1 CO rH rH e— 1 H o H "5 2 H <: en < s o I- M J3 >H O re -^ o < S o CD o r-4 00 to iH -H I-H I— < M • rH . . . . rH • • • m 'J' r-t rH ■ C^ . . . . rH • . . -* f-4 iH . eo • iH rH • • r-i . . . rH • rH • • o 1— ( (N rH rH . I-H rH rH • rH 00 I-H T-H • r^ • • r^ • CO O I-H I-H I-H C-l C^ T-l r-^ (N 'M • (M • • rH • rH : : : : ?) ■ rH . 05 rH 05 r^ rH »i rH ■ M I-H I-H • 1-^ • r-" • •^ in 00 • rH c s] . . ,-( rH rH rH • (N CO ■ • ■ rH CO CO • I-H • IN • (M I-H M • 1-^ CO CO rH • rH C ■ i-t r-t •-* Oi • • in IM (M • I-H iH ■ fH • CO • CO rH CO m CO • • r-i -r-t • I-H • • • CO iH • r^ ' r-t Ol • • C ^ . . , • (M ■ • ■ 05 o "o o o CO p _l i; •^ C C - * 1 ■i « 5 t ~ C C C » ; 3 r — f -' ■H r ^ • * If -H r- > CO ■1 r-t I-H 30 OS O -• r- C^ -H W CO -* I N (M CN C-l C o to r- OS c M O-l 0-1 CM CO rH IM CO -^ CO CO CO CO o H APPENDIX B 149 PQ M Q Oh Sh ^ g CO u ;i£) z u w !?; H /?; H w ;?; C/) [x] 55 '^ t— ( g •^ < i-i W < ;^ LO a; y, < o e i u ] • t^ r^ (M IM • r 00 ^ ,-( ^ (M . . ic • iH •t^rHCOtNiH L n t^ 1-4 r-l C<1 • • rM rH » CO rH « (N IM • i-H 1 M ■^ to .^ . • - T^ T-^ . • CO rH <0 00 i-H i-t in (MrHC^ • T-^ r-* -^ •rHCOCW ■ ' -H •0 -* CQ (N iH • CO (N (M rH • • -«i CI e Tp rH IM • ^ PQ o ii u O o •a '[« C a; Z >4-l w o Z I- w r^ ^^ . (M rH r-t a (M . r^ iH . • r-t rH rH CJ t^ 1-H i-H -M . 1-^ . r-l rH m i-H CO r^ ■^ r^ r-t • r-t Oi 1— t rH rH rH r-l M 00 CO 1-4 r-i • r-\ rH IM !>- CO • T— t • 1-i tH . rH • (M . . la r-t . 1-i • • r-t f^i ' in "* • • rH • rH • . ; H r l< H r r H C H T 1 e K <1 (N 152 LATIN IN SECONDARY SCHOOLS o ^ H H 1 — I H-1 n (-H o W < Zj en u 5 CO o CO • • — 05 00 1^ ' ' • . 1— I • . in • * (N . . . . • ■ rH • rH • • i- H • • CO ^ • O r-t • ' ,-^ . . ' r^ T—< r^ (>] CO f-H < IH rH • 1-i H C-l • rH r-\ t- >H (M ' 1— t r-* • r-i . Ol • (> 1— t • rH OS i—t I-H i-l • I-H • r-t ' - , CN rH • CO o I-H • r-i • * • (>J r-H I-H »M CO 05 . . r^ . ^ ! H ^ . "" ® ;^ 9 5 i^ 8 S a, ^ ^ H 1 1 V. ^ pi; PPEN S FOI Q w H 3 2^ < H K z H^ h^ w S; 0^ o §^ H Ph -u «-3 ;?: < (J S H S « f >^ '-~i H h4 H 1— ( S pq s < ID I o I 00 I I to 50 CO I o to I to in I I to I to o I to I to in T-l I I to . lO • • ■ C^ • ' r^'i r^ . f-t C^l rH T-t tH i-H CO IC C-1 C-1 (M W CO Ml OtOt^OOOSOi-*'MCO'*intOI^050rH-] • •'M -rH .i-H -rHC-J • 'rHrH •C^,-t • CO 5-1 • • (M rH ' !-• i-( 'CO rHC^pH , r^ 1-t ^^ •'MCOrH • .i— I^CO 'l-HX (M • in i-H t-H • r-l ri CO •tOOlOM'fiarHlOC'ltOi-l ■ C) « ■* CI to . T-^ Oi CC .-< C-1 rH CI I-H in r-i(MCO'fiio-oi~050'MeO'«^ H ii ^ 2 e J2 H C o < s fi c 'S o o rH 1 to 1 05 01 1 50 00 10 00 1 CO 00 1 CD 1 r-t 1 to to 10 rH to rH M . e-i r-< N • 00 to 1 to 10 r^ ff-1 rH • M rH CO rH . CO • 1- • rH to rH 10 10 1 1-4 in rH rH rH rH « IN rH tH . • • r- e^ rH to ic ti 1^ ' CO • CO N rH rH . rH CJ y~* in rH in 1 rH CO rH N ^3 . CI r^ CO CO CO rH . rH n rH ?1 CI rH C) CO 1 to CO C^l ■«*' rH to rH • r-t 1-i C-J rH toinrHinci'^rHwc^tM rH rH CO to 00 to 10 =? CO (M S-l • (N • (M rH C-1 05 CO IM • r^ . r~{ Cl 1-t CO CO to • CI rH • ^^ T^ r^ C4 • CO • 00 -^ r^ f~< 1- H • C^ in rH • M ca 00 CO .0 1 • (N • to * rH 1 • rH 1-i • 10 • y-l 00 in rH 1 rH • C^ • 1-i • rH hH rH 1 to . »H . r^ in 1 c ■ in in "o u r H e -1 e Q •> V U -. to I^ c 5 = r- 5 S H r H r T H r H r 1 ^ >--i H ►—( h4 1— 1 m < o I 05 I CO 00 00 I o 00 I I CO o CO CO in I o 13 I CO I CO o n I CO I CO i-H I I CO I CO CO rH IM (N I •* ] m 'CO COiM-^ •OOCC'MCM'>J|-JC-l«C>JfN(N^ ►--1 H 1— 1 J 1— ( pq < < w O o I o in o I 00 o 00 o 10 I CO o CO o 3 c/5 \4 o o I o o> I to 00 00 I o 00 I I to o i"iocot^ocoOr-icic»5'>jiinto«^a>Or-i'* t-lr-(rHi-(i-HrHi-lrHi-tC^C>JC-l!Mi rHrH*..rH.... 05 10 00 r f— t 00 IM • ■ ■ in . • ! I ! I I •CO?-! iH • . . ' r^ rH CO rH 00 r IS •rH . .W • • 1-^ tH • • . '.'.'.'. rH •'••'■'■'•'■'• rH '■'• '• t~ in r '.'.'. '.la '.'•'.'.'.'• '• . rH rH . . . .N . .|M .rHCO • C^ rH 00 rH r to rH IM • . rH CO . • i-t . rH • . . rH . ■ • . rH IN CO rH 10 to f to rHrH..-«I.!!I.! . m C^ . . . . IM . . . i rH . CO iH N IM IN to r to 10 •rH • 1-i r^ > yA .»HC>J • • • in IM • rH ■ ■« • '. -in • ■ • rH CI rH 00 in in r rH 10 rH C^l • • 1-* . . -COlMrHrH • to rH (M • • CO rH . rH ««l . . • (N N N 10 to in r to •(N -T-t ■ •« -rHrH -in •OOINCJ -rHCOrH .rH«> • - irHCON IN HJI in T-t • • ' rA rH ' • rH ©^ CO iH • i-HC:>->"; m rH • -^ W • • • • IM IM CO ■* • • ■ to rH m IN L c u r< IM CO ■* ir: to I- 05 c 01 CO ■* T~i rH rH rH intct-ooo50rHiNeo-*intot-osOrHMi rHr-rHrHrH(MIMOJlM(M3q(M(M w z c <- g3 •2 1^ (^ J2 <-5 < o H h. .2 '35 o o rH OJ ;;;;;;;;;;;;;;;;;;;■;; m ■ o l-O r i-H OS fH-^r-i !!!!! to o l" to 00 in 00 l' I-l 00 ■■■r-l....r-lrH...(M(M!!!!!r-(.ii!l!!! 00 ;---r-(........Mf-H-'i'^'r-.;i;!!rHrH CO o CO r C/J ^1 u Q h^ ^ •-^1 H 1— I J l-H PQ < u S o U o o I o o r 1— t o o> l' to 00 CO I CO to to to f to o to I o o CO r to I to I o o in o . (M Tjl . pH ■ rl CO • • rH O o rH!MCO'*»nt0 1^050'MCO'*L':tOt^GCOiOrH(>JCO-ICOCOCO O 1 62 LATIN IX SECONDARY SCHOOLS h3 a O B3 ;« O pa i GO M o c Pi .2 < w g I O lew ^ o •—I "7? d « pq H < c _o 3 .^ C .0 ■4-t tj I. -4-> n 6 o 00 fH to 13 rH rH rH 1-1 IN • IN CO •■J' . ■^ 1— ( «o ■N 00 r-t lO rH CO rH o rH O rH ■* r-t T-l 05 rH to rH 05 C-1 CO t-t CO rH 00 rH O) (N 00 rH rH rH to IN to l^ IN rH l"I>rH-.)ieq(NearHrHrHO5{<3 1 C 1 C 5 r C H C P 5 C 3C 5 I > C 5 9 "a *-* o APPENDIX D 165 - t) H t— * -^ c '■B ■*-> tn C4 0} 00 00 •* ■^ eo rH •^ r-i 00 CO in i-H rM rH CO r^ ^H t~ rH n r~( 1-1 rH rH CO »H rH »] ^■^ rH rH 00 rH 00 rH rH eo hH r^ rH T-^ i~i r^ OJ 1-t T-l f-H rH N rH !> to • 0^ rH 00 rH rH CO rH U r 1 ^ ) C 5 -fl C U" 5 « 5^ - (X D 1 c > r- H r- 4 C- 1 C< H I- 5 -< ■i 1- f If H r- H r - a H r- 30 H r- < 0- 5 r- H 0- 1 e> 1 5 -a ^ If J e 5 a a 0- 5 t- i ■ a 1 o> i c 1 e« > r- 5 » ^ 0- 3 C< 5 C J H N ir 5 C ! ) 164 LATIN IN SECONDARY SCHOOLS >-• u o U a P^ .2 < ^ ^ "^ ^ .2 W c O < a CO Q l-H OS iH • <-* • rH i-t • rH iH • r-i rH • f-i r~i Hjl CO T— t '■^ r^ C-l 1-i CI 00 in »-H I-H rH CO 1-* rH ^- rH J-t (N CC ^t ^ r^ ■1 rH Tj< r-t t-- r-" rH rH CO rH ■- T-i C] (M 05 T~i • iH T-^ 01 CO ft rH r-t CO rH CO CI CO rH CI rH r-t rH 1-* C^l 'M ^ rr r-< 1—t l^ r-^ CI CO C] 1^ i-H rH C-1 rH CD CM rH ■CO e-1 T»i f-H d !-> r-t 00 CJ CO f-t rH CO -- rH rH rH (>5 r- ri -- ■^ T— ir; r^ - CI J-* CO 10 iH • r^ . • rH rH • (M "^ • r^ rH CO CO CI rH r-t CI •* iH . rH CO • rH r-* • rH rH rH • C^l CO • CO C) CO C-J CI rH CO CO CO • ^^ . • 10 CO • . CD CO rH rH 1-^ r-t rH rH CI CI rH CO (M f— 1 rH • I-H • y~i • m CO • (M CO CI CI • I-H CI rH r-* IM ■* • • rH C>1 • c-l • IM • rH • in 7-i (M • r-i • in • " *-»-^ H 1) a ra H < - to Q Z ^ > .2 S H^ Pi ^ a < CO r-" .2 u 1-1 c: U OS OC r- CC CO CO CO CO c i66 LATIN IN SECONDARY SCHOOLS H O 1^ a < O I o u u < < o > u O J2 J .2 en •a u O o * ; 00 ;?. .'rH ....'.. . (M C^ CO rH..CO--..rH lO in CO ■^ ■^ • • ■ ■>)< •* i-l..ia..CO.rHrH...rH. (M • • -t^ • i~< r^ . 1-t . lyi f~i . . . CO r-i rH • • • Ci • rH .... rH •• C-1 ■ CO rH O rH .rHlO -rHrH -(MrHrHrH .r-i ■ lO r-i 2 1 • rH i-H to • rH lO . ■ rH ■ 0^ ■ ■ . CO i • . ...... r-i OO 1 e<5-;iN.rH..rHrHN..rH. rH r-i S 1 :::"::: :^'^^' ::: : 00 M 1 : : ^'^'^ : : : : I*" : I-' : o rH LO 1 .(NI-HhJ, . . . . rt ' ' ' r-i 1-i * . r~i r^ r-< 'lO ' CO • rH ■ . . . |>) rH rH rH . C-) rH -r-i o rH 21 22 ••-1 . .rH . .rH -rH -M .rH • t~ rH«..f-(....,-|...^^ «3 o ; • • • ■ .rHM . . .rHrHrH • CO I-" • • • r^ -rH . • • r-t r-t r-i * • CO T-i r-t <» oi 1- CO '. '. '. '. ; '"' ■ ^' ■ ■ • O^ r-i r-l • t^ r~t '^r-i • • 'r-i '(M . . .rHrHrHrH 00 r-4 ; i-l ■ ■ -rHeqcMrH t~ O rH •r-i • 'N ' r-i r-i . • • C^ r-i r-i r-i o rH Oi •(M • .C^rH • -rii • • -rHrHrH • rH CO •....rH.t~...rH..rH O rH t^ •r-i • -rHrH • ■* • ,-1 . rH rH rH (M | CO rH CS • ■ • • rn'rH •r-i • • • r-i r-i rH ■^ r-i • C^ • • • W C^ rH 00 1 Hj* • H* O rH • -^ • . . rH rH rH 00 o o u rH c>]r( J3 1 — * C 3 H :^ < u. hJ 3 rt > H H 1— ( hJ ^—1 ■4-» PQ ^ < •0 I- > 10 CO CO (D (M • CO • '^ ; CD 00 '. 1^ • IM • 01 • '. '. rH • • rH to CO • in ■ rH ' • rH rH r-l r-l 1— 1 ■ eo • • IM rH rH I— < 00 •* -* t-» • rH • • CO rH CO (M l-H 1-H CO rH I-H I-H rH •^ rH ' rHCA rH (M ■^ • (N I-H 1-- IM rH . . r- I-H ■'f T-H ■M CO rH (M • • Oi I-H rH (N r-4 I-H CO • rH I-H I-H I-H ^5 I-H I-H rH CO I-H rH -* • • 1—1 1— t IM rH ; • • OirH rH r-l I- a i-H • rH iH I-H • rH rH rH 1^ ^ CO I • rH • IM • • CO 10 TO (M rH • rH • rH I-H C^ 00 CO I-H rH rH rH • 'CO 1- CO CO ; ; • rH rH • ; C4 CO « rH rH rH « eO '-' CO ; C-1 rH I-H rH I^ C C ^ -^ t- - OC 1 C^ 2 > i68 LATIN IN SECONDARY SCHOOLS CJ u u 6 3 in D O pq i$ < CJ O O Lh > E 3 H :<< < u. J O *4-i 12; 1— 1 ^ 3 a) >H H H HH u hJ o 1— ( ■*-• pq 3 < •U O o ; ; ; ; ; rH ; ■ ■ ; I-H \ \ '. CO . rH • • • • H* 00 1-1 i-H • r-l • 1— < l-H I— ( • • CO • W 1-1 rH CO « • ■ • o> rl C^ CO • • (M CO (M CO CO l-H in rM • >-^ ■ • l-H -N • in •*)< -^ G<1 * .-( . rH i-t • •M •* C) ■ CO * 1— t (M r-1 • C'l CO at eg CO rH i ! • ^ CA o iH l-H 1— t * rH CO rH • ) • '. '. ■O" CO (M - ;-i u. s d < en •a u ID m ^ < u t. > ►—1 s > t-H H ^ < < l> W hJ ^ >^ ^ (U HH >H H H 1— I hJ u HH +-» pa n < Q I-. in • • • rH 1-^ 01 1 ■* 1 • r-t . 1-H (N 00 ■* ^^ "^ '^ • r-1 • ' r-t )i ■* rH----rHeO IC •r-t •r-ii~tf-tr-ir^i~ti-ir-tf-t 'r-t rH r-t 0^ T-t • T-t r-t -r-tr-t • ■ (^ r^ r-l • rH r-t • -r-t • -r-t •r-trHrHWOq • • o> • ' r-t r-t r^ r-t -f-lCOiMrH^ • • rH OS •rHrH • -r-t ■ -(MCOrH ■ . . OJ CO CO rH rH - r-t tM r-i r^ r-t - t^ CO •r-tr-t r-t • • • ' r-t •w to CO - • - -r-t r-t lti*~ CO , . . . r-i . . .r-t -r-tr-t • • •rtt CO 1 CO CO r^ • • r^ < - - • »4 e-1 CO '. CO r-t CO 01 00 0^ CO • ■ • • CO to CO IM i-t 01 00 T-t in (-H CO rH i-H rH OS 00 t^ to la •<* CO Oi iH C/3 rH(N'ocO"a