wf/ o' %;>^-^^ -o ***** /•^.\ «^i& % /*£&*% rO "++# v-cr C u J* A A . o > 0° ^ J <* ^ *W .0 ^ .N 1 * *> V v y%. A A 4 %■ '•' ^ i*> .'\ 4>* % ^ ^ ^ A Digitized by the Internet Archive in 2011 with funding from The Library of Congress http://www.archive.org/details/studiesofelement01gray Ube xaniverstts ot Cbtcago STUDIES OF ELEMENTARY- SCHOOL READING THROUGH STAND- ARDIZED TESTS A DISSERTATION SUBMITTED TO THE FACULTY OF THE GRADUATE SCHOOL OF ARTS AND LITERATURE IN CANDIDACY FOR THE DEGREE OF DOCTOR OF PHILOSOPHY (department of education) BY WILLIAM SCOTT GRAY A Private Edition Distributed By The University of Chicago Libraries A Trade Edition is Published By THE UNIVERSITY OF CHICAGO PRESS 1917 1573 Copyright 1917 By The University of Chicago All Rights Reserved Published February 1917 Gtf* Composed and Printed By The University o£ Chicago Press Chicago, Illinois, U.S.A. TABLE OF CONTENTS CHAPTER PAGE I. The Problem i II. Summary of Previous Investigations 2 Description and Criticism of Four Standard Tests 2 Topical Summary of Methods and Results of Reading Investigations 10 Descriptive Bibliography of Reading Investigations 26 III. Description of the Reading Tests 32 The Standard Oral-Reading Test . 32 The Standard Silent-Reading Tests 47 Directions for Giving the Oral-Reading Test 55 Directions for Giving the Silent-Reading Tests 58 IV. Validity of the Oral-Reading Test . 60 Descriptive Summary of the Initial Study 60 Organization of the Second Series of Paragraphs 73 Organization and Standardization of the Third Series of Paragraphs 74 Description of the Measuring Scale for Oral Reading .... 92 Validity of the Adopted Values in the Measuring Scale for Oral Reading 93 Limitations of the Scale 96 V. Validity of the Silent-Reading Tests and Standards ... 99 The Preliminary Study of Silent Reading 99 Critical Discussion of the Adopted Method of Testing Silent Reading 103 Validity of the Standards of Achievement in Speed and Quality of Silent Reading 112 VI. The Investigation of Reading in a City System .... 118 Importance of Reading Investigations 118 Description of the School Population Tested 119 Method of Training Those Who Gave the Tests 120 Summary of the Oral-Reading Results by Grades 120 Variations in Schools 124 iii iv TABLE OF CONTENTS PAGE Analysis to Show Influence of Sex, Method, and Nationality on Progress in Oral Reading 126 Achievement in Cleveland as Compared with Other Schools . . . 128 Results of the Silent-Reading Tests 130 General Relation between Rate and Quality in Silent Reading . . 131 Explanation of the Cleveland Record 138 Summary and Conclusions 141 VII. Special Problems in Reading 142 Periods of Growth in Reading Achievement 142 Comparison of Rates in Oral Reading and in Silent Reading . . 147 Relation of Oral-Reading Achievement to Comprehension of Subject-Matter Read 149 Conclusions 154 LIST OF DIAGRAMS DIAGRAM PAGE I. Progress of 2,193 Pupils in Oral Reading 45 53 54 70 72 83 II. Progress of 2,654 Pupils in Rate of Silent Reading III. Progress of 2,654 Pupils in Quality of Silent Reading IV. "Normal Surface of Frequency" V. Steps of Difference in Difficulty between Paragraphs VI. "Surface of Normal Distribution" VII. Location of the Paragraphs of the Scale on the Base Line of the Probability Curve for Each Grade 87 VIII. Progress of 2,654 Pupils in Rate of Silent Reading . . . . 114 LX. Progress of 2,654 Pupils in Quality of Silent Reading • • • JI 7 X. Graphical Representation of Achievement of Schools in Oral Reading 122 XL Average Scores in Oral Reading in Each Grade in All Elementary Schools and in Four Selected Schools 124 XII. Average Scores in Oral Reading in Each Grade in All Elementary Schools and in Three Selected Schools 125 XIII. Average Scores in Oral Reading for Girls and for Boys in Each of the Eight Grades 127 XIV. Average Scores in Oral Reading of Pupils in Each Grade Using the Ward Method and of Those Using the Aldine Method . 128 XV. Average Scores in Oral Reading of Pupils of Different Nation- alities 129 XVI. Average Scores in Oral Reading in Six Grades in the Cleveland Schools and in Twenty-three Illinois Schools 130 XVII. Average Scores in the Rate of Silent Reading of 1,831 Cleveland Pupils and of 2,654 Pupils of Thirteen Other Cities . . . 135 XVIII. Average Scores in the Quality of Silent Reading of 1,831 Cleve- land Pupils and of 2,654 Pupils of Thirteen Other Cities . . 135 vi LIST OF DIAGRAMS DIAGRAM PAGE XIX. Percentage of 1,831 Cleveland Pupils Found in Each of Nine Groups in Speed and Quality in Silent Reading . . . . 136 XX. Average Scores in Speed and Quality of Silent Reading in Each Grade in Cleveland, in Thirteen Other Cities, and in Three Selected Cleveland Schools 139 XXI. Curve of Progress for Oral Reading . . 143 XXII. Curve of Progress for Rate of Silent Reading 145 XXIII. Curve of Progress for Quality of Silent Reading . . . . . 146 XXIV. Comparison of Oral- and Silent-Reading Rates of 471 Pupils and Students in the University of Chicago 148 XXV. Influence of a Fifteen-Minute Study-Period on Achievement in Oral Reading 151 XXVI. Comprehension of the Subject-Matter Read as Related to the Mastery of Mechanical Difficulties • 152 LIST OF TABLES TABLE PAGE I. Teacher's Check Card (Phonetic) 3 II. Number of Words Read per Second by Grades .... 19 III. Percentage of Increase in Reading Rate through Practice . 20 IV. Comparison of Rates in Oral Reading and in Silent Reading 23 V. Comparison of Ranks of Pupils in Speyer School ... 37 VI. Tabulation Sheet for Individual Records 38 VII. Score Sheet for Individuals and Classes 42 VIII. A Measure for Oral Reading 42 IX. Distribution of the Rankings of the Specimen Paragraphs by Each of the Judges 62 X. Number of Times Each Paragraph Was Given Certain Rankings by the Judges 63 XL Steps of Difference between the Successive Paragraphs of the Scale 64 XII. Percentages Correct and P.E. Equivalents for Each Para- graph 69 XIII. Number of Failures by Grades 81 XIV. Percentage of Failures by Grades 82 XV. Values of the Normal Probability Integral Corresponding to Values of P.E 85 XVI. Location of Each Paragraph of the Scale for Each Grade . 86 XVII. Steps of Difference from One Paragraph to the Next for Each Grade 89 XVIII. Steps of Difference in Difficulty between the Paragraphs of the Scale 90 XIX. Average Shift of the Scale in Difficulty from Grade to Grade 90 XX. Steps of Difference in Difficulty between the Paragraphs of the Scale 91 vii viii LIST OF TABLES TABLE PAGE XXI. Average Shift from Grade to Grade ....... 91 XXII. Measuring Scale for Oral Reading 92 XXIII. Influence of Methods of Scoring on Total Score .... 94 XXIV. Influence of Methods of Scoring on Relative Rank ... 94 XXV. Influence of Giving One-Half Credit to Pupils Failing on Paragraph 1 95 XXVI. Influence of Methods of Scoring on Relative Ranks of Schools 96 XXVII. Speed and Quality Records for 3,161 Preliminary Tests. . 102 XXVIII. Influence of Different Selections on Rank in Silent Reading 106 XXIX. Relative Achievement of Pupils in Reproductions and Answers to Questions 107 XXX. Comparison of the Percentage Reproduced Orally with Per- centage Reproduced in Writing 107 XXXI. Comparison of Reproduction Grades When Scored by the "Word-Counting" Method and by the "Group-Idea" Method in XXXII. Relative Difficulty of the Questions for Each Selection . . 112 XXXIII. Distribution of 2,654 Pupils with Reference to the Time Required to Read 100 Words Silently 113 XXXIV. Distribution of 2,654 Pupils with Reference to Scores for Quality of Silent Reading 116 XXXV. Scores in Oral Reading in Each Grade in Forty-four Schools 121 XXXVI. Specimen Record Sheet Showing Number and Character of Errors Made 126 XXXVII. Distribution of 1,831 Pupils of Cleveland with Reference to the Time Required to Read 100 Words Silently . . . 132 XXXVIII. Average Speed and Quality Scores in Silent Reading of 1,831 Cleveland Pupils and of 2,654 Pupils of Other Cities 134 XXXIX. Average Quality Scores in Silent Reading of All Pupils Reading at Various Rates 137 CHAPTER I THE PROBLEM The purpose of this monograph is to present the results of an investigation made to determine the achievement of boys and girls in oral reading and in silent reading. In the study of oral reading achievement was based on the rate of reading and on the number of errors of various types which were made. In the study of silent reading achievement was based on the rate of reading and on the ability of the pupils to reproduce what was read and to answer specific questions concerning the subject-matter of the tests. The pursuance of this investigation has involved the additional problem of organizing and standardizing tests for use in measuring reading achievement. In this connection an attempt was made to derive reading tests which would secure data that were productive along the following lines : (a) in presenting clearly certain character- istic differences between oral and silent reading; (b) in comparing the achievement of individuals, classes, schools, and entire systems in oral reading and in silent reading; (c) in revealing sources of strength and weakness in the achievement of pupils; (d) in showing the influence of certain factors on achievement, such as methods of teaching, sex, nationality, etc.; (e) in determining the correlation between significant phases of reading, such as speed and comprehen- sion in silent reading. The report of this study divides itself naturally into the follow- ing parts: a summary of previous investigations; a description of the tests used in the investigation; a discussion of the derivation and validity of the oral-reading test and of the silent-reading tests ; a study of reading in a city system; and a report on special prob- lems in reading. CHAPTER II SUMMARY OF PREVIOUS INVESTIGATIONS Up to the present time more than forty investigations have been made which relate to the measurement of achievement in reading. Many of these investigations have been carried out along similar lines, and they can readily be reviewed as a unit. Some' of the investigations have resulted in the development of more or less highly standardized tests of reading. For the purpose of this sum- mary four of these studies, including the tests of Thorndike, Kelly, and Jones, will be discussed separately. In summarizing the investigations which have been made, three purposes have been kept in mind: (i) to give a brief description and criticism of four types of tests which have been devised; (2) to summarize by topics the methods and results of the studies which have been carried on up to the present; and (3) to present a descriptive bibliography to which investigators interested in the problem of measuring reading achievement may refer. DESCRIPTION AND CRITICISM OF FOUR STANDARD TESTS STANDARDS IN MECHANICS OF ELEMENTARY READING This test was devised by R. G. Jones (20:37),* of Rockford, Illinois, to test a pupil's ability to recognize words at sight. It consists of a series of words which appear most frequently in ten widely used primers. A section of the Teacher's Check Card is given below. The words listed in the table are the words upon which the pupil is tested. The numbers to the left of the words indicate the " recurrent values," or the frequency of occurrence of these words in the primers which were analyzed. The numbers in the horizontal line at the top of the table refer to the pupils tested. The entries (X) in the table indicate in each case that the word at the left was missed by the pupil. When the words of the test were originally selected, they were carefully analyzed into sight and phonic words. "In the phonetic 1 See pp. 26 ff. for all such references. SUMMARY OF PREVIOUS INVESTIGATIONS list were tabulated all of those simple English words that belong to the 150 common phonetic families. These are mostly words of a single syllable. In the sight list were tabulated various irregular words that do not employ one of the common phonograms, and most words of two or more syllables" (20:38). After the tabula- tion was complete the words were arranged in the order of their frequency of occurrence in the ten primers analyzed. In giving the tests an investigator may use the two standard vocabulary lists and make separate tests or he may choose words from both lists and make a single composite test. The test is given as follows: A card on which the words are printed in primer type is placed in the hands of the pupil. The teacher has at hand a card similar to the one shown in Table I. As the pupil pronounces the TABLE I Teacher's Check Card (Phonetic) Pupil 1 2 3 4 5 6 7 8 9 10 1,101 bit X X X X X X X X 831 hand 731 stay 622 ran 611 ill X X X X 541 pin 489 cow 433 that 426 hot 403 shake X X X X 395 ten 391 ball 359 gun 315 dig 309 like 288 corn X X X X X X X X X X words of his list the teacher checks on her card the words upon which the pupil fails. The column then presents a fair picture of the pupil's ability to recognize the words which were listed. When all the columns are filled, the card presents a picture of the ability of a class. Thus individual ability, together with the effectiveness or ineffectiveness of class teaching, can be quickly determined. The score of a pupil is determined by finding the sum of the recurrent 4 STUDIES OF ELEMENTARY-SCHOOL READING values of all the words pronounced. The class score is determined by finding the average of the individual scores. The test as it has just been described measures the ability of pupils to recognize individual words at sight. It is open to the objection that the words in the list are not in their natural setting. A pupil may not recognize an isolated word, but will recognize it at once in context. In order to meet this objection, Jones organized sentences in which these words are used and has printed them as reading lessons. As the pupil reads these lessons the teacher checks the errors on her score card. Individual and class scores are determined as described above. As the test is organized it serves as an excellent device for measuring a pupil's ability to recognize simple words at sight, either in isolation or in context. A question arises concerning the validity of a method of calculation which accepts the recurrent values of these words as an index of their difficulty. It would be in harmony with pedagogical principles to find some of the more diffi- cult words occurring in our primers more frequently than some of the less difficult words in order to give the pupil a larger number of opportunities to master them. Whether or not difficulty of pro- nunciation is complicated by factors other than frequency of recurrence in primers was not discussed in detail. It remains therefore for Jones to establish the validity of his assumption. Tkorndike's (43:209) Reading Scale A for visual vocabulary. — The purpose of this scale is to measure the ability of a pupil to know the meanings of single words seen. It was designed for pupils of Grades IV-VIII inclusive, but may be used in the third grade, and is useful in the high school to some extent. A reproduction of the test is given below. Thorndike Reading Scale A VISUAL VOCABULARY Write your name here Look at each word and write the letter F under every word that means a flower. Then look at each word again and write the letter A under every word that means an animal. Then look at each word again and write the letter N under every word that means a boy's name. SUMMARY OF PREVIOUS INVESTIGATIONS 5 Then look at each word again and write the letter G under every word that means a game. Then look at each word again and write the letter B under every word that means a book. Then look at each word again and write the letter T under every word like now or then that means something to do with time. Then look at each word again and write the word GOOD under every word that means something good to be or do. Then look at each word again and write the word BAD under every word that means something bad to be or do. 4. camel, samuel, kind, lily, cruel. 5. cowardly, dominoes, kangaroo, pansy, tennis. 6. during, generous, later, modest, rhinoceros. 7. claude, courteous, isaiah, merciful, reasonable. 8. chrysanthemum, considerate, lynx, prevaricate, reuben. 9. ezra, ichabod, ledger, parchesi, preceding. 10. crocus, dahlia, jonquil, opossum, poltroon. 10.5. begonia, equitable, pretentious, renegade, reprobate. 11. armadillo, iguana, philanthropic. The test is given as follows: A preliminary test similar to Scale A is given by way of preparation. This enables the teacher to make sure that each pupil understands exactly what is to be done. Copies of Scale A are then given out with these instructions: "Read what it says. Do what it tells you to do. Do your best. Hand in your paper when you have marked all the words that you know" (43 '.219). Each pupil is allowed sufficient time to complete the test. When the pupil hands in his paper, it is quickly examined to determine if any of the exercises have been omitted. If the paper is incomplete, the pupil is asked to finish the test. The papers are scored by determining the number of wrongly marked and omitted words in each line. "The highest numbered line that a pupil does with one (or no) omission or error is taken as his score or measure. Similarly, the highest numbered line that the average of a class does with not over one omission or error is taken as the measure for the class" (43:208). From the standpoint of objectivity, accuracy, and precision the scale has been carefully planned. It is easily given and a large number of pupils may be tested at the same time. The records can be scored by relatively untrained testers if necessary. The scale can be increased by alternate series, so that pupils may be tested 6 STUDIES OF ELEMENTARY-SCHOOL READING a number of times without danger of giving special training on the scale words used. The scale measures ability to understand printed words unconfused with ability to express one's self orally or in writing. All of these points are in its favor. Thorndike calls attention to the following limitations of the scale: (i) the small number and variety of words of each degree of difficulty; (2) the omission of pronouns, conjunctions, prepositions, auxiliary verbs, and other words expressing relation; (3) the possi- bility that the scale measures a mixture of general stupidity, indo- lence, or mischief with inability to understand words; (4) the fact that a pupil has one chance in eight of being right by a mere guess. As previously stated, the test measures the pupil's ability to understand printed words only well enough to classify them. A fundamental question arises concerning the practical value of such measure. Thorndike maintains that as long as the degree of under- standing which is measured is objectively denned it makes almost no difference what degree of understanding is measured. If the test is a scientific scale per se, the point of view is probably valid. If the test is to be of greatest value for practical schoolroom pur- poses, it should measure achievement along the line of some clearly denned and highly desirable school product. If we do not know at present just what degree of understanding could be measured with greatest profit, further investigation and analysis should be undertaken to determine the degree of understanding most fre- quently required in the various grades. Thorndike' s {44:445) Scale Alpha 2 for measuring the under- standing of sentences .—The, purpose of this scale is to measure the ability of pupils to understand the meaning of sentences and para- graphs. As prepared for use in the grades, the scale consists of a series of seven carefully graded exercises. The two simplest exer- cises are presented below. Scale Alpha 2. Foe Measuring the Understanding of Sentences set i. difficulty 4 (approximately) Read this and then write the answers. Read it again if you need to. John had two brothers who were both tall. Their names were Will and Fred. John's sister, who was short, was named Mary. John liked Fred SUMMARY OF PREVIOUS INVESTIGATIONS 7 better than either of the others. All of these children except Will had red hair. He had brown hair. 1. Was John's sister tall or short ? 2. How many brothers had John ? 3. What was his sister's name ? SET H. DIFFICULTY 5.25 Read this and then write the answers. Read it again if you need to. Long after the sun had set Tom was still waiting for Jim and Dick to come. "If they do not come before nine o'clock," he said to himself, "I will go on to Boston alone." At half past eight they came, bringing two other boys with them. Tom was very glad to see them and gave each of them one of the apples he had kept. They ate these and he ate one too. Then all went on down the road. 1. When did Jim and Dick come ? 2. What did they do after eating the apples ? 3. Who else came besides Jim and Dick ? 4. How long did Tom say he would wait for them ? The scale is printed in two sheets, one containing Sets I-IV inclusive, the other Sets IV-VII inclusive. In measuring the ability of a class the more difficult sets are not given except in the higher grades. The tests are given in order, beginning with the easiest. Forty minutes are allowed for the test, though most pupils finish well within that time. Hence the time required to perform the test is not considered in scoring the results. A care- fully devised method of scoring the results is presented and tables are included for determining the exact numerical statement of the reading ability of the class. In general, the scale fulfils the stand- ards of objectivity, definiteness, exactitude, and convenience. Scale Alpha 2 measures the ability of pupils to understand the meaning of sentences and paragraphs. Hence it measures one of the most valuable results sought by the school. The specific type of reading ability which is tested is careful analytic reading, such as accompanies problem-solving or grammatical analysis. Hence the test measures only a limited phase of one's ability to get meaning from sentences, and it emphasizes a type of mental ability which is more characteristic of older children than of younger. The time element is not considered; hence one of the important factors which make for efficiency in reading or study is omitted. Inasmuch 8 STUDIES OF ELEMENTARY-SCHOOL READING as the scale consists of a synthesis of various elements, such as harder and harder paragraphs and questions of varying degrees of difficulty, the results can be used only as a gross measure of a pupil's ability to understand sentences. As Thorndike clearly admits, it would be much more valuable if the scale were so devised that it would reveal the specific causes of strength or weakness in the individual's gross achievement. In spite of the limitations which have just been pointed out, this test secures a body of objec- tive data which are very instructive to the classroom teacher, and the results may serve as the basis for intelligent criticism of instruc- tion along certain general lines. The Kansas silent-reading test. — This test was developed and standardized by Dean F. J. Kelly (25:63), of the School of Educa- tion, University of Kansas, while director of the training school in the State Normal School at Emporia. The entire test consists of carefully graded groups of exercises, one for the primary grades, one for the grammar grades, and one for the high school. The following exercises are chosen from the sixteen exercises which form the test for Grades III, IV, and V: No. 2. Value 1.2 Think of the thickness of the peelings of apples and oranges. Put a line around the name of the fruit having the thinner peeling. APPLES ORANGES No. 5. Value 1.6 If you would rather have a dollar than a little stone, do not put a line under dollar, but if you would rather have five dollars than a pencil, put a line under stone. DOLLAR STONE No. 16. Value 8.9 If in the following words e comes right after a more times than e comes just after i, then put a line under each word containing an e and an i; but if e comes just before a more often than right after i, then put a line under each word containing an a and an e. RECEIVE FEATHER TEACHER BELIEVE At the time the test is given the teacher and pupils study the instructions together by way of preparation. A type exercise is given in the instructions. It is presented and discussed in such a SUMMARY OF PREVIOUS INVESTIGATIONS g way that the pupils should know exactly what is to be done. When the teacher is confident that all the pupils understand the instruc- tions, the signal is given to begin. At the end of five minutes the test is discontinued and the papers are collected. The results are scored as follows: "Every answer given is counted either as wholly right or wholly wrong. Where the child's answer is incorrect, cross out the value indicated for that exercise in the margin. Add the values of the exercises which are correctly answered. The sum is the child's score" (26:20). The correct answers to some of the more difficult exercises are given as a safeguard in scoring. This test has been used widely in various parts of the country and it has been highly recommended by many. Professor E. A. Turner, director of the training school of the Illinois State Normal University, who has used the test in his school, made the following comments concerning it in a recent bulletin: "This test will appeal to practical school men. It is definite, simple, and easily presented. The results can be quickly and definitely determined. In practica- bility it ranks with the Thorndike and Ayres handwriting scales . . . ." (45:28). These are very commendable points and their value should not be depreciated. The principles which guided Kelly in the organization of his tests were in substance as follows : The ability to get meaning from the printed page is a complex ability, dependent upon many different factors. The two main factors, each one dependent upon a variety of causes, are the speed with which the reader can get over the lines and the accuracy with which he can comprehend the meaning. In the interest of simplicity it was desirable to combine these two factors in a single mark. The plan was adopted, therefore, of measuring the child's ability to read by the number of reading exercises which he could comprehend accurately in a given time. In order that ability to get meaning might not be confused with ability to reproduce meaning in a written composition, the written interpretation of the exercise was reduced to a minimum. Further- more, Kelly endeavored to devise a test which would test ability to get meaning from the printed page, unconfused with the diffi- culty caused by obscure words or the lack of a particular fund of information. io STUDIES OF ELEMENTARY -SCHOOL READING A careful study of the exercises reveals the fact that the diffi- culties of the selections are determined by a combination of various factors, even though Kelly attempted to avoid this difficulty. The exercises vary widely as to vocabulary and construction. In some cases the exercise contains very few words and in other cases the task imposed on memory is much greater. Some exercises require a direct, straightforward type of reasoning; other exercises are comparable to a puzzle. If the scale were to be one of reasoning through reading, it would seem better to reduce difficulties of vocabulary and construction to a common level and have the diffi- culties in reasoning increase by a series of equal increments. If, on the other hand, it is to be purely a reading test, it would seem that the difficulties in reasoning should be kept more on a common level and those of vocabulary and construction, or both, should increase by a series of equal increments. Furthermore, the test as devised confuses ability to comprehend the problem which is suggested in a given exercise and ability to solve the problem. Evidence comes from many sources that pupils fail to put down the right answer, although they are able at the time to state the character of the answer desired. According to the method of scoring which is used, each correct answer indicates perfect comprehension, while all incorrect answers are interpreted as indicating no comprehension. This interpretation, as we have shown above, is unjust in many cases. It would seem, therefore, that in trying to avoid the confusion of comprehension and repro- duction Kelly has introduced another confusion the significance of which cannot be overlooked. TOPICAL SUMMARY OF METHODS AND RESULTS OF READING INVESTIGATIONS Most of the reading investigations which are summarized under this heading have been carried on with large groups of children under ordinary classroom conditions. Some of them may be criti- cized from the standpoint of scientific accuracy. They are included in this summary because of the suggestiveness of the methods which were employed. A number of laboratory investigations of read- ing are included for the same reason. Each investigation which SUMMARY OF PREVIOUS INVESTIGATIONS n is reported merits separate discussion. The topical method of summarizing these studies has been chosen because of the added emphasis which such a summary gives to certain important issues connected with methods of testing reading. Subject-matter for tests. — The subject-matter for reading tests has been chosen from sources unfamiliar to the pupils and from familiar sources, and, in some cases, it has been composed by the experimenter to meet the special needs of the test. Waldo (47 : 13) selected reading material from a school paper. This was given to the pupils for the first time in the test itself in order that the material might be unfamiliar to all. Starch (42:9), on the other hand, selected passages from grade readers. He justifies this pro- cedure on the ground that passages are then chosen which are best adapted to the reading capacity of children. He further states that "there is every indication on the basis of the tests made thus far that the familiarity gained in ordinary reading in school with any passage will not appreciably affect the tests unless the reading has been done very recently, say within a month preceding the test" (42 : 11). The writer has found that pupils who have read a selec- tion but once recognize it immediately as one that they have read after an interval of one year. Before re-reading the story many pupils were able to reproduce with considerable accuracy its essen- tial points. These facts have led the writer to favor the plan of selecting unfamiliar passages for the test. Gilliland (13 : 13) com- posed several passages to meet certain requirements of his tests. This plan has many advantages, especially if the composer is a fluent writer. Number of selections used. — Investigators differ in their practice relative to the number of selections used in a test series. Courtis (7 : 376) used the same passage for all grades from the fourth through the high school and university. This method makes it possible to use a constant body of material upon which to base rate and repro- duction scores. The method is open to the criticism that the pas- sage is not equally well adapted to the reading capacity of all the students tested. Starch (42 : 2) selected a graded series of passages as test material. He considers that it is highly important to adapt the reading material to the vocabularies and the range of ideas of 12 STUDIES OF ELEMENTARY-SCHOOL READING the respective groups of pupils tested. This he does by selecting a passage appropriate for each grade. Oberholtzer (28:319-22) gave separate selections to each grade. He did not check the diffi- culty of the passages in any way. He gave each grade two tests with passages which were appropriate for the grade. The average results were used as speed records. The validity of this method depends largely upon the appropriateness of the selections. Waldo (48: 255) used a test series consisting of two selections, one of which was used for the third and fourth grades and the other for the fifth, sixth, seventh, and eighth grades. Waldo failed to have either the fourth or the fifth grade read both selections. Hence it was impos- sible for him to compare the achievement of the pupils in the lower grades with the achievement of the pupils in the upper grades. Form in which the material was presented to the pupils. — Many investigators presented the material to the pupils in whatever form it happened to be printed when selected, while other investigators have modified the form somewhat in order to secure more satisfac- tory results. Waldo (47 : 13) used copies of Current Events, a school paper. He distributed the papers face down on each desk . At the signal to begin the papers were turned over and the pupils began reading at an assigned point. Oberholtzer (29:3) used selections chosen from school readers. Each pupil opened his book to the page preceding the one on which the class was to read. At a signal each pupil turned the page of his book and began to read immedi- ately- Brown (5:481) used printed selections of about a page and a half in length so arranged as to have the form of the two pages of an open book. When the printed selections were passed out, they were placed face down on the pupils' desks. At a signal each pupil turned his paper and began to read. Courtis (8:47) used specially printed materials. None of the methods described above takes into account the fact that the average reader must read a few lines before he assumes his habitual rate. This can be easily provided for in the case of Oberholtzer's test by having the class read together a portion of the preceding page before turning to the test passage. Duration of test. — Some investigators who carried on their inves- tigations under laboratory conditions used a definite amount of SUMMARY OF PREVIOUS INVESTIGATIONS 13 subject-matter and made a record of the time required to complete the reading. This method has the advantage that there is a con- stant body of material upon which to base records of speed and of comprehension. The method has a serious limitation in the fact that large groups of pupils cannot be tested at the same time. Most investigators have required pupils to read for a definite period of time, marking the amount read during the reading inter- val. This method enables the investigator to secure records from large groups of pupils at the same time. The objection to the method has been offered that some pupils may not record accurately the amount of subject-matter read. Furthermore, the number of ideas read by the various pupils differs, and as a result it is very difficult to devise an equitable basis upon which to grade repro- ductions. Among the investigators who have used time as the basis for determining the length of the reading test, practice differs as to the length of the reading interval. Starch (42 : 7) had the pupils read for thirty seconds. He supports this practice by citing check tests, which showed that the average amount read per second was prac- tically the same whether the reader read for thirty seconds or for a longer period of time. The subjects in the check tests were adults. His results would have been more convincing had they been secured from children. The majority of investigators have adopted one minute as a convenient reading interval. The work of Courtis (8:47) and of Brown (5:481) may be cited as illustra- tions. Oberholtzer (29:3) adopted a reading interval of two minutes, and Waldo (47 : 13) adopted an interval of five minutes. The objection to the longer periods is a very practical one. As the quantity of reading material increases, we may naturally expect that the amount reproduced will increase. The longer the repro- ductions the more arduous becomes the task of scoring the results. Means of indicating the portion of the selection read. — The investi- gators who have the pupils read a given amount of material are not confronted with this problem. The times at which the reader begins and ends his reading are noted and the total interval is entered on the record sheet. Various methods have been used to indicate the amount read when the pupils read for a definite interval 14 STUDIES OF ELEMENTARY-SCHOOL READING of time. Courtis (8:47) and others had the pupils draw a line around the last word read. Oberholtzer (29:3) had the pupils write down the last word read. This method might lead to some errors, since the same word may appear in each of several lines. This objection can be overcome easily by having pupils record the last sentence or portion of the last sentence read. There is a dis- tinct advantage in having pupils write down rather than mark the last words read. In case the same printed selections are to be used by other pupils, the marks left by previous groups may influence the pupil unduly in the reading rate which he assumes and in the position at which he places his mark. Who shall give the tests. — Abell (1:283) had the students secure their own time records before coming to class. This method allows the widest range of conditions under which the test is given. Very careful instructions may enable mature students to eliminate most of the errors. It is questionable whether results secured in this way by elementary-school pupils are worth much. Waldo (47 : 13) had the same teacher give all the tests in the various rooms, thus eliminating those errors caused by differences in the people giving the tests. Oberholtzer (29 : 2) and others, as Courtis and Starch, had the teachers of the various rooms give the tests to their respec- tive groups of pupils. This method has the advantage that the normal conditions of the classroom are not disturbed as they would be if a stranger were introduced. A source of error lies in the fact that many teachers are untrained in the technique of giving tests. It is pointed out in the report of the Cleveland Survey (23:256) that each teacher who gives tests for the first time finds herself somewhat confused and liable to error. If a sufficient number of preliminary practice tests are given by the teacher, this source of error can be largely eliminated. Directions given to the pupils concerning speed and comprehen- sion. — There is common agreement among investigators that the pupils should be told that they will be questioned concerning the subject-matter read. Practice differs, however, relative to the information given to pupils concerning speed. Mead (27 : 346) says nothing concerning speed, but tells the pupils that when they have finished the reading they will be asked to write all they can remem- ber of what was read. Oberholtzer (29:3) gives practically the SUMMARY OF PREVIOUS INVESTIGATIONS 15 same instructions, telling the pupils that they should read the pas- sages in such a way that they will be able to report to him at the end of the reading. Brown (5:481) tells the pupils that they will be asked to tell what they have read, but that it is desired that they read as much as they can and get the thought. The practice of Brown, all other things being equal, is likely to secure the fairest results. To stress either thought or speed to the exclusion of the other may lead to undue emphasis of one or the other phase of reading. To emphasize both phases in the instructions is likely to lead to more equal distribution of effort during the reading. Terms in which rate is expressed. — Most investigators have expressed the rate in terms of the number of words read per minute or per second of time. Starch (42:7) expressed rate in terms of the number of words read per second. Courtis (8:51) expressed rate in terms of the number of words read per minute. Pintner (32:335) used the number of lines read per minute as the unit for rate. Since the lines in different books vary more in length than do the words appropriate for a given grade, it seems much better, if com- parisons are to be made, to express rate in terms of the number of words read per minute or per second. Average rates may be secured by finding the average number of words read in a given unit of time or by finding the average amount of time required to read a given passage or selection. The average in each case may be expressed in terms of the number of words read per second. Attention is called to the fact at this point that erroneous conclusions will be reached if one compares an average rate secured by one method with an average rate secured by the other method. To illustrate this point, the follow- ing data are presented : Pupil No. of Words Read per Second No. of Words Read per Minute No. of Seconds per 100 Words A B C D i.66| 2.00 2-331 2.66I 80 100 120 140 160 75 60 50 42.8 37-5 E Totals Averages No. of words pe r second 600 120 2.00 265-3 53 -°6 1.86+ 16 STUDIES OF ELEMENTARY-SCHOOL READING The table shows the number of words read per second by each of five pupils; also the number of words read per minute and the number of seconds required to read ioo words. The average results expressed in terms of the number of words read per second differ. Hence if two groups of pupils of equal ability were tested and the average rate for one group were found by one method and the average rate for the second group were found by the other method, the results would differ, and any comparison based on these results would lead to erroneous conclusions concerning the relative ability of the two groups. The two mean or average results described in the foregoing illustration are known respectively as the arithmetic mean and the harmonic mean. Yule (An Introduction to the Theory of Statistics, p. 129) points out the fact that the har- monic mean always gives the smaller average. Hence investiga- tors, when making rigid comparisons of average rates in reading, should select average rates which have been derived by the same method. Methods of testing comprehension. — Practically all investigators have tested comprehension by having pupils write a reproduction of what they have read as soon as the reading has been completed. Pintner (32:335) and Oberholtzer (29:3) are good illustrations. Waldo (47 : 14) used answers to questions in addition to a reproduc- tion. Most of his comparisons are based on answers to questions alone. Kelly (25:64) raises objections to the reproduction method in the following terms: "It is generally agreed, I think, that the ability to reproduce is quite a separate ability from the ability to get meaning, and, therefore, it seems advisable to have a test of the ability to get meaning which involves a minimum of reproduc- tion." The position taken in this monograph is that the mental processes involved in getting meaning are varied and complex. Each of the tests mentioned above measures certain phases of com- prehension. A great variety of methods may and should be used in a thoroughgoing test of one's ability to get meaning. Neither a reproduction test, nor answers to questions, nor Kelly's test, nor all combined, serve as a complete test of comprehension. Methods of scoring comprehension — answers to questions. — Waldo (47:14) prepared a set of ten questions covering two columns of SUMMARY OF PREVIOUS INVESTIGATIONS 17 reading material. If the pupil read all of the subject-matter covered by the ten questions, he received a grade of 10 per cent for each question answered correctly. If the pupil read the subject- matter covered by five questions, a grade of 20 per cent was given for each question answered correctly. The question arises con- cerning the validity of ranking a person who answers three out of five questions right on a basis of equality with another person who answers six out of ten questions right. Scoring reproductions.— Most investigators have scored repro- ductions by the "idea-counting" method. The analysis of a passage into ideas is illustrated by Freeman (22:371) as follows: In our town | when a boy had a coon | he had to have a box turned upside down I to keep it in | and he had to have a little door in the box | to pull the coon out through | whenever he wanted to show it to other boys | or look at it himself | which was forty or fifty times a day | when he first got it. The plan involves determining the total number of ideas cor- rectly reproduced by the reader and expressing this number in terms of the ratio of the reproduced ideas to the total number of ideas. One of the variations from this general procedure is brought out in the following reference. Brown (5:481-82) examined each written reproduction care- fully and determined the percentage of ideas represented. In ascertaining the number of ideas reproduced by each child, every idea was counted which was in most respects complete and in general correctly stated, even though some of the less important details were lacking. The percentage grade determined in this way applied to "quantity of reproduction." The reproductions were examined a second time, and only those ideas were counted which were entirely correct in every respect and of which every detail was reproduced, though not necessarily in the words of the original. The percentage grade determined in this way applied to "quality of reproduction." Comparisons were made on the basis of both quantity of reproduction and quality of reproduction. The objection to this method of scoring reproductions is that it is very difficult to determine whether or not an idea is correctly enough reproduced to be counted as correct. If carefully done, the task is a laborious one. Several of our leading investigators have 1 8 STUDIES OF ELEMENTARY-SCHOOL READING been forced to abandon the method because of its difficulties. The most direct printed statement to this effect is one by Courtis (8 : 50). "The scoring of the amount of material reproduced proved long and tedious, and very few data on this point are available." Waldo (47:15) scores the reproductions by counting the total number of words. He criticizes this method as follows : "In many cases the record of the number of words reproduced means prac- tically nothing regarding the ability of the child in comprehension, for some wrote many meaningless phrases and jumbles of words. Often the phrase written related to other subjects called to mind in some way by the reading material. Others wrote briefly, telling excellently in a few words what was read." Starch (42:7) describes his method of scoring reproductions as follows: "The comprehension is determined by counting the num- ber of words written which correctly reproduce the thought. The written account is carefully read, and all words which either repro- duce the ideas of the test passage incorrectly or add ideas not in the test passage or repeat ideas previously recorded are crossed out." This method has the advantage of being much more rapidly applied than the method of correct ideas. Furthermore, as shown elsewhere in this book, Starch's method is less subjective and almost as accurate as are the more painstaking methods. Index of reading achievement. — Some investigators have expressed achievement in both speed and comprehension in terms of a single index of achievement. Brown (5:482) defines "the reading unit" as follows: "One unit of reading efficiency is a reading ability in which such a rate of reading in words per second is combined with such a power of reproduction that the product of the number rep- resenting the rate and the average of the numbers representing the percentages of quantity and quality is unity. A person who can read at the rate of o . 5 words per second and whose quantity of reproduction is 3 per cent and whose quality of reproduction is 1 per cent has one unit of reading efficiency." Waldo (47:17) multiplied the percentage grade by the number of words read per minute in order to derive a single unit of measurement. Gilliland (13: 19), after finding the number of ideas reproduced, divided this number by the time in seconds required to read the paragraph, SUMMARY OF PREVIOUS INVESTIGATIONS 19 giving the percentage of an idea gained per second. Thus, if 5 ideas were reproduced from a paragraph and it took 15 seconds to read the selection, the "reading value" would be expressed as 33^ per cent. Theoretically, all of these methods assume that the acquisition of ideas proceeds at a uniform rate throughout the reading interval. In reality we know that this does not occur. Nevertheless, practical considerations justify the acceptance of some index which shows general achievement. Rate of silent reading. — It is difficult to draw any final conclu- sions concerning the rate of silent reading from the studies which have been made because of the lack of uniformity in methods of procedure and in the reading materials used. The results of Ober- holtzer (29:7), Courtis (8:56), Brown (5:484), and Starch (42:15) are presented in Table II. Brown used separate selections for the TABLE II Number of Words Read per Second by Grades Grades I II III IV V VI VII VIII Oberholtzer 2-3 2.6 2.6 3-i 3° 3-9 3-6 3-17 3-2 4-7 4-i 4.8 4-6 Courtis Brown 2.01 2.1 Starch i-5 1.8 2.4 2.8 3-6 4.0 third and sixth grades. Oberholtzer used separate selections for each grade without determining the relative difficulty of these selec- tions. Starch chose selections on the assumption that the rate should progress by equal steps from grade to grade. Courtis used the same selection for all grades. In order to make his results readily comparable with the results of other investigators, Courtis' results are expressed in this table in terms of the number of words read per second. In spite of the fact that the methods of testing varied in many details, there is surprisingly close correspondence in the results for the respective grades, particularly so for the third grade. The figures in the table reveal the further fact that progress in rate of silent reading continues throughout the grades. In this connection 20 STUDIES OF ELEMENTARY-SCHOOL READING Courtis (8:51-52) called attention to the fact that the steady progress through the grades reaches adult speed at the ninth or tenth grade and afterward does not vary widely from these con- stant values. He suggests that a value of approximately 320 words per minute, or 5.3 words per second, for rapid reading and 200 words per minute, or 3.3 words per second, for careful reading represent adult ability and the general limit of productive training. Courtis (7:387) gives the following tentative conclusions concern- ing the rate of normal and careful reading: "It will be seen that the curve for normal reading rises to a high value and does not reach its maximum until the high-school years. The curve for careful reading, on the other hand, is practically constant from the sixth grade on. This probably means that the rate and character of one's serious reading are fixed in the early school life." Waldo (47:47) carried on an investigation which showed the importance of the lower grades as a period in which pupils make marked progress in the speed of silent reading. A test was given in the fall and again in the spring to all grades from the third to the eighth inclusive to determine the percentage of increase in the speed of reading. Table III gives the results. TABLE III Percentage of Increase in Reading Rate through Practice Grade Fall Rate Spring Rate Percentage of Increase Grade Fall Rate Spring Rate Percentage of Increase III IV V 76.4 92.7 113. 149. 1 163.3 I29. 2 95-2 76.1 14-3 VI VII VIII.... 128.0 122.7 147.2 130. 1 142.8 158.9 1.2 16.4 8.0 Waldo checked the relative difficulty of the selections which were used in the fall and spring tests. Hence a fair degree of reliance can be placed on his conclusions: "These figures show that the lower grades are very important in the development of reading, for there were made the greatest gains in the mechanics of reading. The third and fourth grades nearly doubled their rates from Sep- tember to March." Characteristics of slow and rapid readers. — Abell (1 : 286) pointed out that slow readers read a word at a time, while the rapid reader SUMMARY OF PREVIOUS INVESTIGATIONS 21 grasps phrases, clauses, and sometimes even sentences at a glance. Quantz (34:50) stated that the superiority of the rapid reader is shown by the fact that his memory of the substance of his reading is more exact than that of the slow reader; he introduces only two- thirds as many irrelevant thoughts; and the ten slowest readers show almost double the amount of lip movement that the most rapid do. Dearborn (9:118) found "that one who reads rapidly in a given style and class of subject-matter will read somewhat proportionately faster than a slow reader whatever within certain limits the nature of the style and the subject-matter." Factors influencing speed. — a) Care with which one reads: Courtis (8 : 50) and others secured results which show quite clearly that one's rate of reading varies with the attention given to the thought of the selection. b) Familiarity with subject-matter: Dearborn, as quoted in Huey (19: 178), found that a reader saved one-third the total read- ing time the second time he read a given selection. c) Lip movements: Hendricks (18:22) points out that in the first grade there is no appreciable difference between the rates at which lip movers and non-lip movers read; in the fourth grade lip movers read more slowly than non-lip movers; and in the eighth grade lip movers read decidedly more slowly than non-lip movers. Huey (19:175) points out exceptions to this general rule. d) Habits of eye movement: Huey's (19:175) observations led him to the following conclusions: "Each would fall into a reading pace that seemed most natural to him and would then read page after page in exactly the same time Habits of eye movement are doubtless important factors in setting this pace." Dearborn (10:115) supports the same point of view when he says: "It is in the writer's belief clearly indicated by the above experiments that one of the essentials of natural and rapid reading is that the reader's eye should at once be able to acquire a regular and uniform motor habit of reaction for each line." Huey (19: 178) points out excep- tions to this rule. Hence it is impossible to state definitely at present the extent to which there is correlation between speed in reading and regularity of eye movements. 22 STUDIES OF ELEMENTARY-SCHOOL READING e) Power of assimilation: Ruediger (37:61) gives the following conclusions in connection with certain of his studies: "After having eliminated the physiological qualities pertaining to the mechanism of vision we saw that neither visual acuity, retinal sensitivity, nor the horizontal extent of acute vision had any significant correlation with reading rate. The essential factors that determine reading must be looked upon as central rather than peripheral. It is not a matter of getting material to the brain, but of assimilating it after it is there." Ruediger points out that the correlation between the number of reading pauses and the reading rate is apparently at variance with the conclusion just given. He states that the fact that more is seen at a fixation by a rapid reader than by a slow reader may just as well be the result of a strong power of assimila- tion as of more effective vision. /) Length of words: Slightly different points of view are held relative to the influence of the length of words on reading rate. Beer states that the reading time varies with the character of the words used. Thus a preponderance of monosyllabic words makes the reading time longer owing to the relatively greater meaning conveyed by monosyllabic words. Messmer reported similar results. Dearborn (9:98), in checking up Messmer 's experiment, says: "It is not the short words as such, but the words which can- not be easily grouped with others which necessitate separate fixa- tion." . g) Summary of factors: Quantz (34:50), considering all the factors which he found to contribute to rapid reading, said that they are, in the order of importance, "visual perception, practice as determined by amount of reading from childhood on, power of concentration, mental alertness estimated by rapidity of original composition, scholarly ability as decided by college records." Influence of practice on rate. — Huey and Dearborn each early expressed the belief that the reading rate might be increased by practice. Oberholtzer (29:8) reports an increase of 50 per cent in the rate after two weeks of practice in rapid reading. The tests showed also that oral expression and power to grasp the content were equally improved. Bowden (4:41) reports the results of some experiments with adults concerning the influence of practice on rate SUMMARY OF PREVIOUS INVESTIGATIONS 23 as follows: "The figures show that for these two types of subjects the reading rate may be markedly increased by practice " Bowden (4:45) reports the effect of practice upon the rate of silent reading of a seven-year-old boy. The practice exercise in rapid silent reading was a ten-minute one daily for forty days. In the first exercise the boy read twelve words in a minute; in the last he read fifty-five words in a minute. Miss Bowden points out that the case is not a typical one, but she believes that the rapid-reading exercise had no small share of influence in the wonderful improve- ment made by the boy, and that other elementary-school children whose motor habits are not fixed might secure improvement in rate from simple exercises in speed of reading. Pintner (33 : 129) performed a series of experiments to determine whether the process of articulation could be eliminated, and if so, whether by practice the ordinary rate of reading and the average degree of comprehension could be attained. His conclusions are as follows: "From the results I think we are justified in saying that reading without articulation can take place, that it can be accom- plished as quickly as reading with articulation, and that practice in reading without articulation increases the ordinary rate of reading, no doubt due to the fact that after such practice the amount of articulation made use of is not so great as formerly." Pintner's study does not determine whether articulation can be entirely eliminated from reading. Comparison of the rates of oral and silent reading. — Most of the studies made show the superiority of silent reading. Oberholtzer (29: 7) presents a comparison between the rates of oral reading and TABLE IV Comparison of Rates in Oral Reading and in Silent Reading Grade Oral Reading Silent Reading Grade Oral Reading Silent Reading III rv v 2. I 2-3 2.4 2-3 2.6 31 VI VII VIII 2.8 3 1 3-9 3-9 4-7 4.8 silent reading, as shown in Table IV. These figures show clearly that silent reading is a much more rapid process than oral reading 24 STUDIES OF ELEMENTARY-SCHOOL READING and that the increase in the intermediate grades is much more marked in silent than in oral reading. Hendricks (18: 16) found that first-grade pupils read somewhat more rapidly silently than orally. In the fourth and eighth grades the increase was very marked, being almost twice as rapid silently as orally in the eighth grade. His data further show the very interesting fact that if pupils are urged to increase their speed of silent reading in the third grade they become confused and read less rapidly. Mead (27:348) tested six sixth-grade classes and found that all classes but one read more rapidly silently. Pintner (32 : 335) tested fourth-grade children and found that they averaged 20 lines a minute orally and 28 lines a minute silently. Loss in oral-reading rate during the summer vacation. — P. C. Packer and H. W. Anderson (31:81) tested the rate of oral reading of pupils in the Iowa City public schools in Grades IB-VB inclu- sive to determine the influence of the summer vacation upon the pupils' reading rates. The rate of oral reading was determined during the latter part of May, 1915, and again on September 18, 19 1 5, after one week of school. All children read from material with which they were familiar, and were allowed any prose selection they liked. The results showed that the pupils' reading rates had been decidedly lowered: "Grade IB read at the median rate of 50 words per minute in May and 44 words in September, a loss of 6 words per minute. Grade IA fell from 84 words to 49 words; Grade IIB from 125 to 68; Grade IIA from 145 to 124. Grade IIIB lost 14 words in its rate. Grade IIIA lost 18 words, and I VB, IVA, and VB lost respectively 14, 13, and 13 words in rate of reading from May to the end of the first week of school in Sep- tember." Relation between speed and quality. — Various positions have been taken with regard to this problem. Abell (1:284) stated that comprehension may be independent of absolute rate, since of three classes of readers — fast, moderate, and slow — some comprehend well and others fairly or poorly. Quantz (34:30) found that the rapid readers were on the average about 37 per cent superior to the slow readers in the quality of their work. "The superiority of the rapid reader is also shown by the fact that his memory of the SUMMARY OF PREVIOUS INVESTIGATIONS 25 substance of his reading is more exact than that of the slow reader. He introduces only two-thirds as many thoughts not found in the original selections." Waldo (47 : 54) plotted the correlation between speed and comprehension for several grades. His conclusion is as follows: "No definite results can be stated, though it would seem that the rapid readers usually are strong in comprehension, although there are many exceptions." Hendricks (18:6) shows distinct positive correlation between speed and quality of silent reading. "In the percentage of thought reproduced the rapid readers excel, giving 91 per cent of the thought as compared with 76 per cent reproduced by the slow readers." Growth in comprehension. — Because of the wide variations in method and the differences in character of subject-matter used it is difficult to make many comparisons concerning comprehension. Growth in comprehension seems to continue at a more steady rate and to continue longer than growth in speed. Starch (42:15) reports comprehension scores in terms of the number of words written: Grade I II m IV V VI VII vni Words written 15 20 24 28 33 38 45 50 The figures reveal the fact that the progress is steady throughout the grades, the steps of increase averaging slightly more in the upper grades than in the lower. Comparison of comprehension during oral and silent reading. — Mead (27:34) found that reproductions were slightly superior in all sixth-grade classes tested when the reading was done silently. Pintner (32:336) reported that fourth-grade children reproduced 34 per cent of the points when reading orally and reproduced 40 per cent of the points when reading silently. The advantage seems to lie on the side of silent reading. When we take into consideration the greater amount of subject-matter read per unit of time silently, the argument is decidedly in favor of silent reading. 26 STUDIES OF ELEMENTARY-SCHOOL READING DESCRIPTIVE BIBLIOGRAPHY OF READING INVESTIGATIONS The following bibliography contains practically all the class- room investigations which have been published. Such laboratory investigations are included as might prove most suggestive to anyone who wishes to pursue this work further: i. Abell, Adelaide M. "Rapid Reading; Advantages and Methods," Edu- cational Review, VIII (October, 1894), 283-86. Reports the results of an investigation with 41 Wellesley students to determine the relation of the rate of silent reading and the comprehension of subject-matter. Points out important characteristics of slow and rapid readers. 2. Beer, Max. "Die Abhangigkeit der Lesezeit von psychologischen und sprachlichen Faktoren," Z.f. Psychologie, LVI (1910), 264-303. Reports the results of an investigation to determine the influence of length of words upon rate of reading. 3. Boggs, Lucinda Pearl. "How Children Learn to Read: An Experimental Study," Pedagogical Seminary, XII (December, 1905), 496-502. Reports the results of an investigation with four boys under school age to determine relative efficiency of teaching oral reading by the alpha- betic, phonetic, word, and sentence methods. 4. Bowden, Josephine H. "Learning to Read" (Master's thesis, University of Chicago, 1911), Elementary School Teacher, XII (September, 1911), 21-33- Reports the results of an investigation to determine how children recognize words. Two experiments are reported: first, an individual study, and, second, a class study. 5. Brown, H. A. "The Measurement of the Efficiency of Instruction in Reading," Elementary School Teacher, XIV (June, 1914), 477-90. Reports the results of an investigation with "a little over four hundred" third- and sixth-grade pupils in seven school systems to determine the rate of silent reading and the quantity and quality of reproduction. Dis- cusses at length the relation of the measurement of reading ability to reconstruction of methods of teaching reading. 6. Brown, H. A. The Measurement of Ability to Read, Bulletin No. 1, New Hampshire Department of Public Instruction, Concord, New Hampshire. A manual of directions for giving and scoring reading tests and diag- nosing class and individual needs. 7. Courtis, S. A. "Standard Tests in English," Elementary School Teacher, XIV (April, 1914), 374-92. Reports the results of an investigation with pupils chosen from grades above the third, from the high school, and from adults, to determine rates of reading and ability to comprehend what was read. Describes and dis- SUMMARY OF PREVIOUS INVESTIGATIONS 27 cusses each of the six tests included in what the author calls "English Tests." 8. Courtis, S. A. "Standards in Rates of Reading," Fourteenth Yearbook of the National Society for the Study of Education, Part I (191 5), pp. 44-58. Reports the results and conclusions reached in a first attempt to deter- mine standard scores in some of the measurable elements that make up skill in reading. Standards of achievement are reported for pupils in Grades IV- VIII, for high-school pupils, and for adults. 9. Dearborn, W. F. Psychology of Reading (1906), p. 98. In checking up Messmer's experiments concerning the influence of length of words on reading rate Dearborn concluded: "It is not the short words as such, but the words which cannot be easily grouped with others which necessitate separate fixations." Pp. 116-22: Reports the results of an investigation to determine the influence of subject-matter, type, and eye movements upon the rate of reading. 10. Dearborn, W. F. The Psychology of Reading, Columbia University Con- tributions to Philosophy and Psychology, Vol. XIV, No. 1 (1906). 11. Department of Education, the city of New York. Teachers Yearbook of Educational Investigations, No. 6 (191 5), pp. 43-45. Reports the tests which have been devised for measuring the efficiency of instruction in reading. The work of Brown, Judd, Thorndike, and Gray are mentioned. 12. Gill, E. J. "Methods of Teaching Reading," Journal of Experimental Pedagogy, I (March, 191 2), 243-48. Reports results of investigation with young pupils to determine whether the phonic or sentence method of teaching reading accomplishes the best results. 13. Gilliland, A. R. "A Comparative Study of Oral and Silent Reading" (Master's thesis, University of Ohio, 191 5). Also Journal of Educational Psychology, VII ^April, 1916), 201-12. Reports the results of an investigation with 120 subjects chosen from the grades above the second, from the high school, and from the University of Ohio to determine the relative rates of oral and silent reading of children and adults at different stages of development, and to determine relative value of oral and silent reading in securing thought. 14. Gray, C. T. "The Relation of Breathing to Oral Reading," Journal of Educational Psychology, IV (January, 19 13), 39-41. Reports the results of laboratory investigations with 50 grade pupils and adults to determine the relation of breathing to oral reading. 15. Gray, William S. "A Tentative Scale for the Measurement of Oral Reading Achievement" (Master's thesis, Columbia University, June, 1914). Reports methods and results of testing oral-reading achievement of grade pupils by means of a series of standardized paragraphs arranged in the order of increasing difficulty. 28 STUDIES OF ELEMENTARY-SCHOOL READING 16. Gray, William S. "Methods of Testing Reading, I," Elementary School Journal, XVI (January, 191 6). Describes certain rough tests of oral and silent reading which can be readily given by any teacher in connection with the regular work of the classroom. 17. Gray, William S. "Methods of Testing Reading, II," Elementary School Journal, XVI (February, 1916), 281-98. Describes standard tests in oral and silent reading which were used in connection with the surveys in Cleveland and Grand Rapids. 18. Hendricks, Eldo L. A Study in Reading (Silver, Burdett & Co., 1911). Reports the results of an investigation to determine for pupils of Grades I A, IV A, and VIII the rate of ordinary silent reading, very rapid silent reading, and ordinary oral reading, and the percentage of thought reproduced in each case. Additional problems concerning the characteris- tics of ordinary silent reading are discussed. 19. Huey, E. B. The Psychology and Pedagogy of Reading (1908), pp. 174-75. Reports the results of an investigation with 20 graduate students to determine rate of reading. Ordinary and maximum rates of oral and silent reading are reported. The influence of lip movements and habits of eye movements are discussed. 20. Jones, R. G. "Standard Vocabulary," Fourteenth Yearbook of the National Society for the Study of Education, Part I (1915), pp. 37-43. Reports the results of tests given to two first-grade classes to determine the ability of pupils to pronounce simple words at sight. The method of devising a vocabulary test is described in some detail. 21. Jones, R. G. Standards in Mechanics of Elementary Reading (Bulletin, Rockford Printing Co., Rockford, Illinois, 1915). Describes tests which will aid in discovering a pupil's command of the vocabulary common to first-, second-, and third-grade readers. 22. Judd, Charles H. "Reading Tests," Elementary School Teacher, XIV (April, 1914), 365-73- Attention is called to two fundamental types of reading — oral and silent. The value of comparisons between groups undertaking similar work is emphasized. Practical suggestions are given concerning the selec- tion of material for reading tests and concerning the administration of tests. 23. Judd, Charles H. Measuring the Work of the Public Schools (The Survey Committee of the Cleveland Foundation, 1916), pp. 124-61. Reports the results of the investigations of reading carried on in con- nection with the Cleveland Survey. 24. Judd, Charles H. "Reading," Fifteenth Yearbook of the National Society for the Study of Education, Part I (1916), pp. 111-19. Reports certain results secured in connection with the Cleveland Survey. The rates of oral reading and silent reading are compared, and the relation between speed and quality of silent reading is discussed. SUMMARY OF PREVIOUS INVESTIGATIONS 29 25. Kelly, J. F. "The Kansas Silent Reading Tests," Journal of Educational Psychology, VII (February, 1916), 63-80. Describes the organization, derivation, and use of a series of tests to measure the pupil's comprehension of what is read. 26. Kelly, J. F. The Kansas Silent Reading Tests, Studies by the Bureau of Educational Measurements and Standards, No. 3 (Kansas State Normal School, Emporia). 27. Mead, Cyrus D. " Silent Reading versus Oral Reading with One Hundred Sixth Grade Pupils," Journal of Educational Psychology, VI (June, 191 5), 345-48. Reports the results of an investigation to determine the relative value of oral reading and silent reading as means of securing thought from the printed page. 28. Oberholtzer, E. E. "Testing the Efficiency in Reading in the Grades,' Elementary School Journal, XV (February, 1915), 313-22. Reports the results of an investigation with 1,800 pupils in Grades I- VIII to determine the speed of oral and of silent reading. 29. Oberholtzer, E.E. "Testing the Efficiency in Reading in the Grades" (Master's thesis, University of Chicago, August, 191 5). Reports the results of an investigation with 1,800 pupils in Grades I- VIII in Tulsa, Oklahoma, to determine (a) rate of oral reading and of silent reading; (b) rate of reading in successive grades; (c) factors of greatest influence in affecting rate; (d) relation of speed and comprehension. 30. Oberholtzer, E. E. "The Effects of Efficiency Tests in Reading on a City School System," Fifteenth Yearbook of the National Society for the Study of Education, Part I (1914), pp. 138-40. Summarizes briefly the values which resulted from giving tests of reading in his school system. 31. Packer, P. C, and Anderson, H. W. "The Loss in Reading Ability during the Summer Vacation," Midland Schools, XXX (November, 1915), 81-82. Reports the results of an investigation with 566 pupils in Grades I-V in Iowa City public schools to determine the effect of the summer vacation upon the rate of oral reading. 32. Pintner, Rudolph. "Oral and Silent Reading of Fourth Grade Children," Journal of Educational Psychology, IV (June, 1913), 333-37. Reports results of investigation with twenty-three fourth-grade pupils to determine whether the percentage of subject-matter retained or com- prehended is increased or decreased by the supposed help of oral reading. 33. Pintner, Rudolph. "Inner Speech," Psychological Review, XX, 129-53. Reports results of experiments to determine whether articulation could be eliminated during silent reading, and if so whether by practice the ordinary rate of reading and the average degree of comprehension could be increased. 30 STUDIES OF ELEMENTARY -SCHOOL READING 34. Quantz, J. A. "Problems in the Psychology of Reading," Psychological Review, Monograph Supplement, Vol. II, No. 1 (1897), pp. 1-51. Reports the results of an investigation with 50 Juniors and Seniors of the University of Wisconsin to determine the normal and maximum rates of oral and silent reading, and to determine some of the factors and condi- tions upon which rate of reading depends. 35. Romanes, G. J. Mental Evolution in Animals (1884), pp. 136-37. Reports the results of an investigation with "practiced readers" to determine the speed of silent reading and the factors conditioning it. Gives certain conclusions concerning the rate of reading and the power of assimilation. 36. Rubins, R. B. "Tests of the Seven Year School" (Master's thesis, Uni- versity of Chicago, September, 191 5), pp. 32-37. Reports the results of Gray's reading tests given to 232 grade pupils in Bristol, Tennessee. Steady progress beyond the third grade is reported. 37. Ruediger, W. E. Field of Distinct Vision, Columbia Contributions to Philosophy and Psychology, Vol. XVII, No. 1 (1907). In connection with laboratory experiments upon the field of distinct vision Ruediger reports results concerning the correlation between the horizontal extent of acute vision and reading rate. Emphasizes the importance of power to assimilate. 38. Salt Lake City Survey. The Test in Reading (June, 1915), pp. 154-65. Reports the results of Courtis reading tests given to 1,624 pupils of Salt Lake City. Achievement in speed of reading, comprehension of subject- matter, and the correlation between speed and memory tests are given. 39. Schmitt, Clara. "School Subjects as Materials for Tests of Mental Ability," Elementary School Journal, XV (November, 1914), 150-61. Reports results of tests given to pupils of Grades I-V and to a group of defective children to determine the relative achievement of each group as to speed and errors in oral reading, together with the ability to reproduce and to interpret the subject-matter read. 40. Sholty, Myrtle. "A Study of the Reading Vocabulary of Children," Elementary School Teacher, XII (February, 1912), 272-77. Reports results of investigation to determine the reading vocabulary of second-grade pupils. Five points were reported: total number of words in reading vocabulary; number of words known in context; number known out of context when seen on instant; number pupils could build up when allowed to see words for some time; number pupils could neither recognize on instant nor build up when given time. 41. Starch, Daniel. The Measurement of Efficiency in Reading, Writing, Spell- ing, and English" (College Book Store, Madison, Wisconsin, 1914). Presents the subject-matter for silent-reading tests for each grade, gives instructions for administering and scoring the tests, and presents standards of achievement for each grade SUMMARY OF PREVIOUS INVESTIGATIONS 31 42. Starch, Daniel. "The Measurement of Efficiency in Reading," Journal of Educational Psychology, VI (January, 191 5), 1-24. Reports the results of tests given to 3,511 pupils in fifteen schools of Wisconsin, Minnesota, and New York to determine speed of silent reading and comprehension of subject-matter. 43. Thorndike, E. L. "The Measurement of Ability to Read," Teachers Col- lege Record, XV (September, 1914), 207-27. Presents preliminary scales and tests for the measurement of achieve- ment in reading along the following lines: (1) a pupil's ability to under- stand the meaning of words and sentences seen so far as concerns (a) the understanding of words singly, and (b) the understanding of sentences and paragraphs; (2) a pupil's ability to pronounce words and sentences seen in a series of paragraphs arranged in the order of increasing difficulty. 44. Thorndike, E. L. "An Improved Scale for Measuring Ability in Read- ing," Teachers College Record, XVI (November, 191 5), 445-67; concluded XVII (January, 191 6), 40-67. Describes the derivation and use of an improved and extended form of the Alpha Reading Scale described in the Teachers College Record for September, 1914. 45. Turner, Edwin A. " Standards Employed in the Determination of Teach- ing Efficiency," Normal School Quarterly, Illinois State Normal University, January, 1916, pp. 26-29. Reports the tests which have been devised for measuring the efficiency of instruction in reading. The studies of Thorndike, Kelly, and Gray are mentioned. 46. Valentine, C. W. "Experiments on Methods of Teaching Reading," Journal of Experimental Pedagogy and Training College Record, II (1913), 99-112. Reviewed by E. H. Cameron, Psychological Bulletin, XI, 329. Reports the result of an investigation of the comparative merits of the word and phonic methods of teaching reading. 47. Waldo, Karl D. "Tests in Reading in Sycamore Schools" (Master's thesis, University of Chicago, July, 19 14). Reports results of an investigation in Grades III- VIII in the public schools of Sycamore, Illinois, to determine the growth in speed and com- prehension of silent reading during a period of six months. 48. Waldo, Karl D. "Tests in Reading in Sycamore Schools," Elementary School Journal, XV (January, 191 5), 251-68. Gives a summary of results presented in a Master's thesis, University of Chicago, 1914, showing the growth made by various grades in speed of silent reading and comprehension of subject-matter during a period of six months. CHAPTER III DESCRIPTION OF THE READING TESTS Tests of oral reading and of silent reading were organized and standardized for use in the investigations which are reported in chaps, vi and vii. The technical details connected with the deriva- tion and standardization of these tests will be discussed in chaps, iv and v. It is the purpose of this chapter to describe only the general features of the tests, to outline the methods of giving them, and to describe how the results are scored. THE STANDARD ORAL-READING TEST The oral-reading test which was used in the investigations reported in this monograph consisted of a series of twelve passages arranged in the order of increasing difficulty. The passages of the test are shown on p. 33 Conditions under which the tests should be given. — The value of the results secured through the use of tests depends largely upon the conditions under which they are given. In view of this fact those who gave the tests were instructed to observe carefully the following points : a) The tests should be given in a well-lighted room, preferably in an office or alcove, or in a room which is free from noises and interruptions. b) The presence of visitors while the tests are being given is undesirable. Experience has shown that the pupil is disturbed very little by the presence of the experimenter, but that the presence of a third person is distracting. If anyone asks to observe the testing, it is well to refuse the request politely but firmly. c) The attitude of the experimenter toward the pupil is a very important factor. Greet the pupil with a casual remark of some kind, thus putting him at his ease. During the progress of the test do not say or do anything which would encourage or discourage 32 DESCRIPTION OF THE READING TESTS 33 A boy had a dog. The dog ran into the woods. The boy ran after the dog. He wanted the dog to go home. But the dog would not go home. The little boy said, "I cannot go home without my dog. " Then the boy began to cry. Once there were a cat and a mouse. They lived in the same house. The cat bit off the mouse's tail. "Pray, puss," said the mouse, "give me my long tail again." "No," said the cat, "I will not give you your tail till you bring me some milk." 3 Once there lived a king and queen in a large palace. But the fting and queen were not happy. There were no little children in the house or gar- den. One day they found a poor little boy and girl at their door. They took them into the palace and made them their own. The king and queen were then happy. Once I went home from the city for a sum- mer's rest. I took my gun for a stroll in the woods where I had shot many squirrels. I put my gun against a tree and lay down upon the leaves. Soon I was fast asleep dreaming of a group of merry, laughing children running and playing about me on all sides. One of the most interesting birds which ever lived in my bird-room was a blue jay named Jakie. He was full of business from morning till night, scarcely ever still. He had been stolen from a nest long before he could fly, and he had been reared in a house long before he had been given to me as a pet. 6 The part of farming enjoyed most by a boy is the making of maple sugar. It is better than blackberrying and almost as good as fishing. One reason why a boy likes this work is that some- one else does most of it. It is a sort of work in which he can appear to be very industrious and yet do but little. 7 It was one of those wonderful evenings such as are found only in this magnificent region. The sun had sunk behind the mountains, but it was still light. The pretty twilight glow embraced a third of the sky, and against its brilliancy stood the dull white masses of the mountains in evident contrast. 8 The crown and glory of a useful life is charac- ter. It is the noblest possession of man. It con- stitutes a rank in itself, an estate in the general good will, dignifying every station and exalting every position in society. It exercises a greater power than wealth, and is a valuable means of securing honor. 9 He was six feet tall and his body was well proportioned. His complexion inclined to the florid; his eyes were blue and remarkably far apart. A profusion of hair covered the forehead. He was scrupulously neat in his appearance; and, although he habitually left his tent early, he was well dressed. 10 Responding to the impulse of habit, Josephus spoke as of old. The others listened attentively but in grim and contemptuous silence. He spoke at length, continuously, persistently, and ingra- tiatingly. Finally exhausted through loss of strength he hesitated. As always happens in such exigencies, he was lost. 11 The attractions of the American prairies as well as of the alluvial deposits of Egypt have been overcome by the azure skies of Italy and the antiquities of Roman architecture. My delight in the antique and my fondness for architectural and archaeological studies verges onto a fanati- cism. 12 The hypotheses concerning physical phenom- ena formulated by the early philosophers proved to be inconsistent and in general not universally applicable. Before relatively accurate principles could be established, physicists, mathematicians, and statisticians had to combine forces and work arduously. 34 STUDIES OF ELEMENTARY -SCHOOL READING the pupil. Give him no hint concerning his progress until the test has been completed. d) Arrangements should be made with the grade teacher to have the pupils sent to the experimenter. In the first three grades the following plan should be used: At the outset two pupils should be sent to the room in which the tests are being given. While the first pupil is being tested, the second pupil should remain at the door outside the room. When the first test has been completed, the pupil should be dismissed and the second pupil should enter for his test. As soon as the first pupil reaches the classroom, a third pupil should be sent to the door of the testroom. In this way the experimenter always has a pupil at hand to be tested as soon as a subject is dismissed. In the grades above the third two pupils should be sent at first and one thereafter at the end of each five minutes. Slight readjustments may be necessary in regard to the time schedule, since some experimenters learn to test more rapidly than others. Directions for giving the oral-reading tests. — i. The purpose of this test is to determine the rate of oral reading and the ability of pupils to pronounce words and sentences at sight. 2. When everything is in readiness to begin, hand the pupil a copy of the paragraphs and give the following directions: " I should like to have you read some of these paragraphs for me. Begin with the first paragraph when I say 'Begin.' Stop at the end of each paragraph until I say 'Next.' If you should find some hard words, read them as best you can without help and continue reading." In case a pupil in one of the first two grades hesitates several seconds on a difficult word, pronounce it for the pupil so that he may con- tinue reading. Mark the word as mispronounced. 3. While the pupil is reading, record two sets of facts in regard to the reading, the time required to read each paragraph and the errors made. a) The time record is secured by noting the exact second at which the pupil begins reading a paragraph and the time when he completes it. The number of seconds required to read the para- graph should be recorded in the margin to the right of the para- graph. DESCRIPTION OF THE READING TESTS 35 b) In order to illustrate clearly the character of errors and the method of .recording them, the following paragraph is inserted: The sun pierced into my* large windows. It was the opening of October, and the sky was^of)a dazzling blue. I looked out of my window (and) down the street. The white houseQof the long, slight street were {almost painful to the eyes. The clear atmosphere allowed full play to jhjj sunjsj?rightness. If a word is wholly mispronounced, underline it, as in the case of "atmosphere." If a portion of a word is mispronounced, mark appropriately, as indicated above: "pierced" pronounced in two syllables, sounding long a in "dazzling," omitting the s in "houses" or the al from "almost" or the r in "straight." Omitted words are marked as in the case of "of" and "and"; substitutions as in the case of "many" for "my"; insertions as in the case of "clear"; and repetitions as in the case of "to the sun's." Two or more words should be repeated to count as a repetition. It is very difficult to record the exact nature of each error. Do this as accurately as you can. In all cases where you are unable to define clearly the specific character of the error, underline the word or portion of the word mispronounced. Be sure that you mark each error. In case you are not sure that an error was made, give the pupil the benefit of your doubt. If the pupil has a slight foreign accent, distinguish carefully between this difficulty and real errors. 4. Each pupil should be allowed to continue reading until he makes at least the following number of errors in each of two para- graphs: 5 errors or more in 40 or more seconds, or 7 or more errors in case the paragraph is read in less than 40 seconds. Scope of the oral-reading test. — A question naturally arises at this point concerning the scope of the test. To what extent do measures of rate and mechanical errors constitute a measure of oral-reading ability ? In the first place, the elementary school must of necessity place considerable emphasis on the mechanics of reading. To the extent, therefore, that the test measures progress along this par- ticular line of school endeavor it becomes a useful tool in checking achievement in oral reading. 36 STUDIES OF ELEMENTARY-SCHOOL READING In the second place, rate in itself is a very important measure of oral-reading ability. When a child first learns to read, he recog- nizes words very slowly. As the association between the sight of the symbol and its pronunciation becomes more firmly established, the pronunciation of the word follows more closely upon its sight. This increase in the rate at which pupils recognize words continues throughout the grades. The following figures give the number of words read per second by large numbers of pupils of each grade when reading the third paragraph of the reading test: Grade I II III IV V VI VII vin Words per second 1.36 2. 21 2.42 2.98 3-39 3-54 363 3-70 Furthermore, as the subject-matter to be read increases in difficulty, the speed of reading is decreased, although no errors may be made. This fact is shown very clearly in the case of pupil No. 1 in Table VI. Because one's reading rate increases through the grades and because one's rate of reading is influenced by the diffi- culty of the subject-matter, it is fair to assume that of two readers the one who recognizes words the more readily has the greater ability in oral reading. Moreover, ability to read rapidly is a fair measure of the mastery which the reader has of the printed page. It was pointed out in the Cleveland monograph, entitled Measuring the Work of the Public Schools, that "the poor reader is one who is unable to pass readily from the printed symbol to the meaning. For the poor reader the mere mechanical processes are obstacles and he loses time in trying to perform the preliminary mental acts which are necessary before he can comprehend the passage. In the case of the good reader, on the other hand, the mechanics of the process are very fluent and rapid. The proficient reader has mastered the words and moves on without hesitation to the meaning." The error in relying entirely on rate as a measure of efficiency in reading was also pointed out in the Cleveland monograph. DESCRIPTION OF THE READING TESTS 37 Again, it has been found that the relative rank in a class which a pupil holds on the basis of the oral-reading test is very closely correlated to the rank given him by his teacher on the basis of general reading ability. Table V presents the ranks of the pupils TABLE V Comparison of Ranks of Pupils in Speyer School Pupil Rank by Teacher Rank by Test Pupil Rank by Teacher Rank by Test WUda Ogden Cornelia . Jack Athea Marion Rosa Jessie Parker I 2 3 4 5 6 7 8 9 I 2 3 4 IO n 5 6 7- Grace Donald . Margaret . . . Hugh Rupert Muriel EUy Reginald . . . LiUan IO II 12 13 14 15 16 17 18 8 9 12 14 13 15 16 17 18 of the sixth grade of the Speyer School, New York City, given in 19 14 by the teachers of that school on the basis of general reading ability, and the ranks given to the same pupils on the basis of the oral- reading test. As revealed by the table, the correlation between achievement as measured by the test and achievement as measured in the classroom is very high indeed. Similar comparisons have been made in several grades of the Elementary School of the Uni- versity of Chicago with corresponding results. The writer, while giving more than two thousand oral-reading tests personally, observed that pupils who ranked high in the oral- reading test usually gave much better expression to their reading than those who ranked low. C. T. Gray, who worked on reading during 19 15- 16 in the laboratory of the School of Education of the University of Chicago, experimented with a number of pupils who had taken the oral-reading test. Those pupils who were sent to him with a low rank according to the test usually received a much lower grade on expression than did those who received a high rank. A specific test for expression was not included in this study for the following reasons: The writer worked for months trying to devise some simple standards by which to judge the quality of expression. No test devised would be accepted as a whole by a 38 STUDIES OF ELEMENTARY-SCHOOL READING single teacher called into conference. It therefore seemed futile, at the present stage of development of scientific measuring, to include such a test in this investigation. Method of tabulating data. — For the purpose of computing indi- vidual and class scores it is desirable to tabulate the records of the pupils in as convenient a form as possible. To this end either of the following methods may be used: If the investigator is interested in the final score only, the total number of errors made in reading a given paragraph may be recorded on the test sheet under the num- ber expressing the time in seconds required to read that paragraph. This may be done at the time the tests are given. The scores can then be computed quickly from the original test sheets by the method described in later paragraphs of this chapter. If the inves- tigator wishes to make a careful study of the causes of high or low records or if he desires to make productive comparisons of the achievement of individuals, classes, or entire schools, a more elaborate method of tabulation is desirable. The following method of tabulation which was adopted for use in this investigation was found to be effective for the purposes of analysis and comparison: The records for each school were entered on a separate sheet. The TABLE VI Tabulation Sheet for Individual Records Pupils Paragraph No. i Paragraph No. 2 Paragraph No. 3 Paragraph No. 4 G N s A T G M s I R T ( J M S ] [ R T G M ( ) S ] r R T G M s I R TT I 2 3 4 5 6 7 G G B G G G B 7 7 8 7 7 8 7 13 IS 12 28 19 27 40 I I I I I I I 2 19. 20 . 20 . 35- 20 . 39- 46. 20 20 37 38 38 55 25 25 35 63 45 69 I I I I I 2 I 3 1 2 1 1 TT I . I . 1 1 2 . I I . . I . I . . 2 . I . 1 . 1 1 . I II TT . 1 1 I T TT I . 1 2 . 3 ? TT . 1 . 2 II 2 2 . records for each grade were tabulated as a unit in order from the first to the eighth. Individual records within a grade were entered in the order of achievement from the best to the poorest. The time required to read each paragraph and the number of errors of each type which were made in reading were recorded. Table VI DESCRIPTION OF THE READING TESTS 39 illustrates the method of tabulation adopted. As presented, it is limited on account of space to the first four paragraphs of the oral- reading test. Description of tabulation sheet for class and individual records. — The letters G, N, S, A under "Pupils" refer to grade, to the rank of the pupil, to sex, and to age, respectively. The letters T, G, M, O, S, I, R under each paragraph refer to the following items: T, to time in seconds ; G, to gross errors; M, to minor errors; O, to omis- sions; S, to substitutions; I, to insertions; and R, to repetitions. The record of the first pupil entered in the table reads as follows : The pupil was in the second grade, ranked highest in the class, and was a girl, seven years old. She read the first four paragraphs of the scale in 13, 19, 20, and 25 seconds, respectively, with no errors. The record of the last pupil entered reads as follows : This second- grade pupil ranked lowest in the class and was a boy seven years old. He read paragraph 1 in 40 seconds with one gross error and two repetitions. He read paragraph 2 in 46 seconds with two minor errors, two omissions, two substitutions, and one repetition. The other paragraphs were too difficult for him to read effectively. Standards for scoring. — The problem of scoring results is much more difficult for reading than it is for many other subjects. In arithmetic the problem is a relatively easy one because standards of perfect accuracy can be adopted as the basis for grading. Such a standard in reading would be open to serious objections, since many of our most effective readers make some errors. The scoring of the record of each individual depends upon the facts that a cer- tain number of errors may reasonably be expected for each para- graph and that a certain amount of time must be allowed for the reading. If the pupil exceeds a given time limit or makes more than a given number of errors, the amount of credit which he gets for reading the passage should be proportionately reduced. Upon the basis of the time required for reading and the number of errors made by large numbers of pupils, four standards of varying degrees of severity were adopted as the basis for scoring. The validity of these standards is discussed on p. 76. They are given below in the order of increasing severity. 40 STUDIES OF ELEMENTARY-SCHOOL READING A record is checked as a failure if: A. It is not read because of previous failure. It is read in 40 or more seconds with five or more errors. It is read in less than 40 seconds with seven or more errors. B. It is not read because of previous failure. It is read in 30 or more seconds with four or more errors. It is read in less than 30 seconds with five or more errors. C. It is not read because of previous failure. It is read in 25 or more seconds with three or more errors. It is read in less than 25 seconds with four or more errors. D. It is not read because of previous failure. It is read in 20 or more seconds with two or more errors. It is read in less than 20 seconds with three or more errors. Individual and class scores. — Individual scores were determined by applying each of the four standards to the pupil's record for each paragraph. If the paragraph was read successfully under the con- ditions of two of the standards, the figure "2" was entered in an appropriate column of the score sheet (Table VII) ; if the reading met the conditions of one, three, or four of the standards, the appro- priate figure was entered on the sheet. As shown in Table VI, pupil No. 7 read paragraph 1 in 40 seconds with three errors. It is apparent that this paragraph was read successfully according to the conditions of Standards A and B. The record does not meet the conditions of Standard C because more than 25 seconds were required to read it and there was a total of three errors. Since the reading was a failure according to Standard C, it was also a failure according to Standard D, which is more diffi- cult. The paragraph was therefore read successfully under two standards. Paragraph 2 was read in 46 seconds with seven errors. An application of Standard A shows that the reading was a failure according to the conditions of the most liberal standard, and hence a failure ' under the conditions of each of the more diffi- cult standards. The paragraph was therefore read successfully under none of the standards. In case a paragraph was not read successfully by any of the standards, no entry was made in the table. DESCRIPTION OF THE READING TESTS 41 In order to secure speed and accuracy in scoring the results, the four standards mentioned above were summarized in the form of the following table: Seconds Errors 7 or More 40 or more 3o-39 2S-29 20-24 19 or less . . The numbers in the horizontal line at the top of the table refer to the number of errors made. The numbers in the left-hand column of the table refer to the number of seconds required to read a paragraph. The entries in the table refer to the number of standards under which a paragraph is read successfully. The table reads as follows : If a paragraph is read in 40 or more seconds with one error, it is read successfully according to four standards; if two errors are made, it is read successfully according to three standards, etc. With the aid of this table it is possible to score individual records with great rapidity. DESCRIPTION OF SCORE SHEET EOR INDIVIDUALS AND CLASSES The letters G, N, S, A under "Pupils" in Table VII refer to the grade, to the rank of the pupil, to sex, and to age, respectively. The numerals under "Paragraph," refer to the paragraphs of the scale in order. The entries in the table refer to the number of standards under which each paragraph was read successfully. The table reads : The first pupil entered in the table was a second-grade pupil, ranked highest in the class, and was a girl, seven years old. She read paragraphs 1, 2, 3, 4, 5, and 6 successfully by all four standards, paragraph 7 by one standard, and the remaining paragraphs by none of the standards. The total class score and the average class score for each paragraph are presented at the foot of the table. The total class score was determined by finding the sum of the scores for each paragraph separately. These sums were 42 STUDIES OF ELEMENTARY-SCHOOL READING then entered on the score sheet at the foot of the appropriate columns. Since the highest possible number of successes per para- graph is four times the number of pupils tested, the average score by paragraphs was found by dividing each item of the total score by four. TABLE VII Score Sheet for Individuals and Classes Pupils Paragraphs G N s A i 2 3 4 s 6 7 8 9 10 ii 12 II... I 2 3 4 5 6 7 G G B G G G B 7 7 8 7 7 8 7 4 4 4 2 4 4 2 4 4 3 4 3 2 4 4 2 3 i i 4 4 i 4 2 2 4 2 I I II... II.. II.... II.... II.... II.... Total class s Average cla.? core 24 6 20 5 IS 3-7 9 2. 2 8 2 6 i-5 2 °-5 s score .... Upon the basis of the type of scoring just outlined the steps of difference in difficulty between the paragraphs and the shifts in difficulty from grade to grade have been determined from data for 3,299 pupils and are represented in Table VIII. The derivation TABLE VIII A Measure for Oral Reading Grade 5 10 IS 20 25 30 35 40 45 So 55 60 65 70 75 80 8S 7 go 95 100 I.... I 5 6 7 8 10 6 7 8 10 11 2 7 8 10 11 12 3 8 10 11 12 5 IO II 12 6 10 11 12 II... I 2 3 5 6 7 2 3 5 6 7 8 2 3 5 6 7 8 3 5 6 7 8 10 III. . 1 2 3 5 6 11 12 12 IV... I 2 3 5 v.... I 2 3 VT I 2 VTT I I VTTT and validity of Table VIII, together with the reasons for the omission of paragraph 4 and paragraph 9 from the table, are discussed in chap. iv. The horizontal line of numbers at the top of the table marks off twenty equal steps between o and 100. These numbers may be DESCRIPTION OF THE READING TESTS 43 interpreted as indicating either the relative difficulty of the para- graphs or the amount of credit which should be given for the suc- cessful reading of the various paragraphs. The Roman numerals to the left of the table refer to the eight grades of the elementary school. The entries in the table refer to the respective paragraphs of the test. The table reads as follows : In the first grade a pupil who reads paragraph 1 successfully should receive a credit of 55 points; for reading paragraph 2 successfully he should receive 10. additional points, or a total of 65 points; for reading paragraph 3 successfully he should receive 5 additional points, or a total of 70 points, etc. The successful reading of paragraph 1 does not repre- sent so great an achievement in the second grade as in the first grade. According to the table, a second-grade pupil receives a credit of 35 points for reading paragraph 1 successfully. Ten addi- tional points, or a total of 45 points, are given for reading para- graph 2 successfully. Five additional points, or a total of 50 points, are given for reading paragraph 3 successfully, etc. The numerical value of the achievement of a given class was de- termined according to the following method: The average class score for each paragraph was first found as indicated on p. 42. The aver- age class scores for the second-grade class mentioned in Table VII were as follows : an average of six successful readings of paragraph 1 ; 5 successful readings of paragraph 2; 3.75 successful readings of paragraph 3; 2.25 successful readings of paragraph 4; 2 successful readings of paragraph 5 ; 1.5 successful readings of paragraph 6 ; o '. 5 successful reading of paragraph 7. According to Table VIII, 35 points of credit should be given for each successful reading of para- graph 1 ; six successful readings amount to a total value of 210. To this should be added 10 additional points for each of 5 successful readings of paragraph 2 ; 5 additional points for each of 3 . 75 success- ful readings of paragraph 3; 5 additional points for each of 2.25 successful readings of paragraph 4 ; 5 additional points for each of 2 successful readings of paragraph 5 ; 5 additional points for each of 1 . 5 successful readings of paragraph 6 ; and 2 . 5 additional points for 0.5 successful reading of paragraph 7. The total value is the sum of 210, 50, 18.75, 11. 2 5> IO > 7 • 5, 2. 5, or 310. Since there were seven members in this class, the average score for the class was one-seventh 44 STUDIES OF ELEMENTARY-SCHOOL READING of 310, or 44+. The foregoing description of the method of find- ing the numerical score for a class has been reduced, with slight modifications, to the series of specific directions which follow: a) Find the sum of the scores for each paragraph separately, and enter each total score at the foot of the appropriate column on the score sheet. b) Enter the total score for each paragraph in the column under "Score," in the following table: Paragraph Score X Value Product I 2 IO 3 5 VALUE FOR PARAGRAPH I 5 5 Grade 1 55 6 5 II 35 7 5 HI 30 8 5 IV 25 10 10 V 20 11 IO VI 15 12 5 VII 10 VIII 5 Total prodi let Average class score c) The value or credit given for the successful reading of paragraph 1 varies with the grade. These values are given in the column to the right of the table. Enter the appropriate value for paragraph 1 in the blank space in the column under "Value." Thus, the appropriate value for paragraph 1, for the third grade, is 30. The values for all other paragraphs remain the same for all grades. d) Multiply the score for each paragraph by its value and enter the result in the column under "Product." e) Find the sum of the products and divide by 4 times the number of pupils in the class. The result is the average class score. Graphical representation of oral-reading scores. — As was stated in preceding paragraphs, the ability to read a certain passage with- out error means less on the part of a pupil in the upper grades than on the part of a pupil in the lower grades. Hence grades are to be compared with each other by the recognition of different levels of achievement. These different levels, as determined from a large number of records, can be expressed graphically as indicated by the vertical lines in Diagram I. Each line represents the scale for a grade and begins below at the point where the score of 10 should DESCRIPTION OF THE READING TESTS 45 be represented. Higher scores can be represented by appropriate distances along the vertical line above 10. In the diagram the lines end at the points where the score of 70 belongs for each grade. The dotted oblique lines above and below, connecting the successive 70's and io's respectively, indicate the curves of progress which Diagram I. — Progress of 2,193 pupils in oral reading would result if, in the one case, all scores were 10, or if, in the other case, all scores were 70. The solid line near the middle of the figure represents the record of 2,193 pupils. This average record constitutes a standard with which individual schools may be com- pared. The numerical value of the record for each grade as repre- sented in the diagram is as follows : First grade, 3 1 ; second grade, 46 STUDIES OF ELEMENTARY-SCHOOL READING Standardized Reading Paragraphs William S. Gray School- Pupil. . Teacher. .Nationality. .Grade. .Age — A boy had a dog. The dog ran into the woods. The boy ran after the dog. He wanted the dog to go home. But the dog would not go home. The little boy said, "I cannot go home without my dog. " Then the boy began to cry. Once there was a little pig. He lived with his mother in a pen. One day he saw his four feet. "Mother," he said, "what can I do with my feet?" His mother said, "You can run with them." So the little pig ran round and round the pen. Once there were a cat and a mouse, i They lived in the same house. The cat bit off the mouse's tail. "Pray, puss," said the mouse, "give me my long tail again. ' ' "No" said the cat, "I will not give you your tail till you bring me some milk." Once there lived a king and queen in a large palace. But the king and queen were not happy. There were no little children in the house or gar- den. One day they found a poor little boy and girl at their door. They took them into the beautiful palace and made them their own. The king and queen were then happy. One of the most interesting birds which ever lived in my bird-room was a blue-jay named Jackie. He was full of business from morning till night, scarcely ever still. He had been stolen from a nest long before he could fly, and he had been reared in a house long before he had been given to me as a pet. 6 The part of farming enjoyed most by a boy is the making of maple sugar. It is better than blackberrying and almost as good as fishing. One reason why a boy likes this work is that someone else does most of it. It is a sort of work in which he can appear to be very industrious and yet do but little. 7 It was one of those wonderful evenings such as are found only in this magnificent region. The sun had sunk behind the mountains, but it was still light. The pretty twilight glow embraced a third of the sky, and against its brilliancy stood the dull white masses of the mountains in evident contrast. 8 The crown and glory of a useful life is char- acter. It is the noblest possession of man. It forms a rank in itself, an estate in the general good will, dignifying every station and exalting every position in society. It exercises a greater power than wealth, and is a valuable means of securing honor. 9 He was approximately six feet tall and his body was well proportioned. His complexion inclined to the florid; his eyes were blue and re- markably far apart. A profusion of hair covered the forehead. He was scrupulously neat in his appearance ; and, although he habitually left his tent early, he was well dressed. 10 Responding to the impulse of habit Josephus spoke as of old. The others listened attentively but in grim and contemptuous silence. He spoke at length, continuously, persistently, and ingra- tiatingly. Finally exhausted through loss of strength he hesitated. As always happens in such exigencies he was lost. 11 The attractions of the American prairies as well as of the alluvial deposits of Egypt have been overcome by the azure skies of Italy and the antiquities of Roman architecture. My delight in the antique and my fondness for architectural and archaeological studies verges onto a fanati- cism. 12 The hypotheses concerning physical phenom- ena formulated by the early philosophers proved to be inconsistent and in general not universally applicable. Before relatively accurate principles could be established, physicists, mathematicians, and statisticians had to combine forces and work arduously. DESCRIPTION OF THE READING TESTS 47 42; third grade, 46; fourth grade, 47; fifth grade, 48; sixth grade, 49; seventh grade, 47; eighth grade, 48. Standardized reading paragraphs. — The oral-reading test as given on p. 33 was used in the investigations which are reported in chaps. vi and vii. During the progress of these investigations it was found that several changes were desirable in connection with cer- tain paragraphs. These modifications are described at the end of chap, iv, and they have been incorporated in the test as it appears on p. 46. The revised form of the oral-reading test is the one which was used in connection with the Grand Rapids Survey. Arrangements are now being made to use it in the St. Louis Survey, in a survey of eleven cities of northern Illinois, and in connection with studies of reading in numerous cities throughout the country. The average oral-reading scores for 4,066 pupils of Grand Rapids are given below: Average Oral-Reading Scores Date Grades I II III IV V VI VII vm Grand Rapids March 1916 35 44 47 49 So 47 48 48 THE STANDARD SILENT-READING TESTS The subject-matter of the silent-reading tests consists of three selections entitled "Tiny Tad," "The Grasshoppers," and "Ancient Ships." These selections were adapted respectively to the interests and reading capacity of second- and third-grade pupils, fourth-, fifth-, and sixth-grade pupils, and seventh- and eighth-grade pupils. The phases of reading ability which are tested by these selections are rate of silent reading, ability to reproduce the subject-matter read, and ability to answer specific questions concerning what was read. That speed is an important element in silent reading cannot be questioned. Of two pupils, all other things being equal, that pupil 48 STUDIES OF ELEMENTARY-SCHOOL READING who reads at the rate of ioo words per minute will achieve only one-half as much in a given unit of time as one who reads at the rate of 200 words per minute. As the diagram on p. 136 shows, the rapid reader usually secures a larger number of ideas from what he reads than does the slow reader. For these reasons it is evident that the facts concerning the rate of silent reading are of the greatest importance to supervisors and teachers. It will be pointed out on p. 105 that when a person tries to understand what he reads, a number of complex processes are involved, and that meaning or comprehension may take various forms. In the present investigation it was impossible to test com- prehension of what was read in all its forms. Practical considera- tions demanded that comprehension be measured along one or two representative lines. It was decided, therefore, to measure the ability of pupils to reproduce what was read, and to measure ability to answer questions concerning specific points. These two measures were selected because comprehension is tested very largely in the classroom by one or both of these methods. Subject-matter of the silent-reading tests. — The selections used for the tests of silent reading were printed on cards so that they could be easily handled. Each selection was printed in three sections. The middle section contained 100 words in the case of the easiest selection and 200 words in the case of each of the two more difficult selections. The section at the left of the middle section serves a double purpose : First, it gives the pupil something to read by way of preparation for the test part of the passage; secondly, the tester can readily determine the moment at which the reader moves his eyes from the bottom of the card to the top of the card where the words upon which the time record is based begin. In the following passages the ends of sections are indicated by the short horizontal lines. The passages used are as follows : Tiny Tad (for second and third grades) Tiny Tad was a queer little fellow with only two legs and a short tail. He was nearly black, too, and much smaller than most tadpoles in the big pond. He could hardly wait for his front legs to grow. DESCRIPTION OF THE READING TESTS 49 "When I have them all," he said, "I'll leave this dirty water and go up into the orchard. What fun it will be to hop and hop and hop. If only I had a little brother to hop with me, I should be so happy." It wasn't long before his legs began to grow. He moved about and kicked aroiind until his legs were quite strong. "I am going out on the bank to see if I can hop," he said one night when he was just six weeks old. The sun was hardly up the next morning when a little toad jumped out of the water and hopped up on the bank. He was very small, but none too small for the little legs that wabbled under him. It was Tiny, the young toad. The Grasshoppers (for fourth, fifth, and sixth grades) The grasshoppers were among the worst enemies of the early settlers of Nebraska. Their homes were on the high plains and among the hills at the foot of the great mountains in the West. Here they lived and raised their families. In dry seasons there were more children and less food at home. Then they assembled and flew away in great swarms to the east and to the south. They traveled hundreds of miles. Sometimes on clear, warm, moonlight nights they traveled all night. More often they settled down late in the afternoon and fed, and then continued their way the next day. The great grasshopper raid took place in September, 1874. Suddenly along the frontier states the air to the west was filled with grasshoppers. There were billions of them in the great clouds which darkened the sun. The noise of their wings filled the air with a roaring sound like a rushing storm, followed by a deep hush as they dropped to the earth and began to devour the crop. All the corn was eaten in a single day. Where cornfields stood at sunrise nothing remained at night but stumps of stalks swarming with hungry hoppers struggling for the last bite. They stripped the garden patches bare. They gnawed great holes in the rugs and carpets put out to save favorite plants. The buds and fruit of trees were consumed. They followed the potatoes and onions into the ground. When they finished the garden and green crops, they attacked the wheat and oats in the shock and the wild grass in the unplowed fields. Only two green crops escaped them, broom corn and sorghum cane. They did not seem to have a sweet tooth. 50 STUDIES OF ELEMENTARY-SCHOOL READING Ancient Ships (for seventh and eighth grades) There is no more interesting study to marine architects than that of the growth of modern ships from their earliest form. Ancient ships of war and of commerce equally interest them; but as they study the sculptures and writings of the ancients, they find records of warships far outnumbering ships of commerce. Among ancient nations the Greeks and Romans were the best shipbuilders. Judging from the description of their works their crafts must have been elegant, swift, and seaworthy. This is more than can be said of many of the more showy productions of the shipyards of Britain, France, and Spain even so late as the Middle Ages. • There is no question now that the ships of the ancients made extended voyages urged by oars alone. A thousand oarsmen were sometimes required to man the sweeps, besides a crew of five hundred soldiers and sailors. Written descriptions give us splendid pictures of fleets of these ancient ships moving swiftly along the white villa-dotted shores of Greece, or majestically sweeping into some mirror-like harbor and with sounding trumpets saluting the setting of the low, western sun. We are able to make from old records very fair models of these ancient warships. One writer describes the great galley of Philopator as propelled by forty banks of oars. His description is questioned, for however plain the description of these warships may be, no one has yet shown the precise manner in which forty banks could be arranged. A bank of oars means a row on one deck, and while there are many pictures of galleys they show nothing more than a trireme, which is a ship of three banks. A ship of forty banks puzzles our imagination. After the pupil had read the selection appropriate for his grade he was given a sheet of printed directions. On one side of the sheet the directions state that the pupil should write thereon all he could remember of the story which he had read. As soon as he completed his reproduction he turned the sheet and answered the ten questions which were printed there. The questions for each of the selections are as follows: Tiny Tad i. How many legs did Tiny Tad have at the first of the story? 2. How did Tiny compare in size with most of the other tadpoles in the pond? DESCRIPTION OF THE READING TESTS 51 3. Which legs did Tiny wish would grow? 4. Where did Tiny say he would go when he got all his legs ? 5. What did Tiny wish he had to hop with him? 6. What did Tiny do to make his legs grow strong ? 7. How old was Tiny when he decided to leave the pond? 8. What part of the morning did Tiny choose for leaving the pond ? 9. How did Tiny get upon the bank ? 10. What size was Tiny at the end of the story ? The Grasshoppers 1. In what western state were the grasshoppers enemies to the settlers? 2. What effect did dry seasons have on the number of grasshoppers? 3. On what kind of nights did the grasshoppers sometimes travel all night long? 4. When the grasshoppers were making their long journeys, what would they often do late in the afternoon ? 5. In what year did the great grasshopper raid take place ? 6. Like what did the great groups of grasshoppers look as they traveled through the air ? 7. What sort of noise did they make when flying through the air? 8. What change was brought about in the appearance of the cornfields by the grasshoppers between morning and night ? 9. What did the settlers do to protect their favorite plants ? 10. Why didn't the grasshoppers eat the broom corn and sorghum ? Ancient Ships 1 . To whom is the study of the growth of modern ships interesting ? 2. How do the records of warships compare in number with the records of the ships of commerce ? 3. What peoples were the best shipbuilders among the ancient nations? 4. How did the ancient vessels compare in elegance and swiftness with the more showy productions of the Middle Ages ? 5. What kind of voyages were sometimes made by ancient ships when propelled by oars only ? 6. What was the total number of men required on some of the ships ? 7. Explain clearly what a "white villa-dotted shore" means to you. 8. From what source do we secure the ideas which enable us to make models of the ancient warships? 9. What does a "bank of oars" mean? 10. Why do we question the statement that the great galley of Philopator had forty banks of oars ? Directions for giving the silent-reading tests. — 1. The purpose of this test is to determine rate and comprehension in silent reading. 52 STUDIES OF ELEMENTARY-SCHOOL READING 2. Grades II and III are tested on "Tiny Tad," Grades IV, V, and VI on "The Grasshoppers," Grades VII and VIII on "Ancient Ships." 3. The teacher first fills in the four blanks at the top of a report sheet: Name School , etc. 4. The teacher then hands the pupil a card with the selection appropriate for his grade printed on it, with these directions : " Read the story on this card silently. Read the story from beginning to end without stopping or repeating any of it. Read the story rapidly but carefully. Do not stop reading to ask about difficult words; read them as well as you can. Be prepared to tell the story or to answer any question about it when you are through. Do you understand ? " Make the directions clear to the pupil before begin- ning the reading. 5. Secure the time record as follows: When the pupil shifts from the bottom of the first column to the top of the middle column, note the position of the second hand. When he shifts from the bottom of the middle column to the top of the third column, note the second hand again. Record the number of seconds required to read the middle column. Be sure that your record indicates the time for reading the middle column only. If possible, use a stop watch. 6. If the pupil has been reading "Tiny Tad," ask him to tell you the story of Tiny Tad as well as he can. In the meantime write as rapidly as possible just what the pupil says. Ask the pupil to tell it a sentence at a time if he talks too rapidly. When he discontinues speaking, ask him if he remembers anything else. Following this, ask him the questions and record his answers. 7. If the pupil has been reading "The Grasshoppers" or "Ancient Ships," hand him the report blank to fill out. Ask him to write the story which he has read as well as he can. Impress upon him the fact that he should not look at the questions on the opposite side of the sheet until he has finished writing his story. While the pupil is filling the blank, the next pupil may take the test. DESCRIPTION OF THE READING TESTS 53 Speed of silent reading. — The speed of silent reading was deter- mined as follows : The average number of seconds required by each grade to read ioo words was determined. This average was then expressed in terms of the number of words read per second by dividing ioo by the average number of seconds required to read ioo words. Diagram II shows the progress of 2,654 pupils in rate of silent reading, representing thirteen city systems, and may be used as a standard for comparison. Since three selections were used in order to adapt the subject-matter to the maturity of pupils 4,00 3} 21 2l97 3.00 2i21 1197 ft.no 1^21 j.97 1.00 2nd 3rd 4th 5th 6 th 7 th Diagram II. — Progress of 2,654 pupils in rate of silent reading 8 th of different grades, a readjustment was necessary in the diagram. The points of this readjustment are between the third and fourth grades and between the sixth and seventh grades. A dotted line is drawn in the diagram dividing the curves of progress at these points. The figures at the left of the diagram indicate the number of words read per second when the easy selection was read; the figures on the line between the third and fourth grades indicate the equivalent number of words read per second when the second selec- tion was used; and the figures on the line between the sixth and seventh grades indicate the equivalent rate for the most difficult 54 STUDIES OF ELEMENTARY-SCHOOL READING selection. The average number of words read per second by each of the grades is as follows: "Tiny Tad" Second Third I.50 2.30 "The Grasshoppers" Fourth Fifth Sixth 2.20 2.57 2.79 "Ancient Ships" Seventh Eighth 2.69 2.87 Quality of silent reading.- — The quality of the pupil's reading was determined as follows: All wrong statements, irrelevant statements, and repetitions were checked from the pupil's reproduction, and the remaining words counted. The ratio of these words to the total number of words forms the reproduction grade. For each question 60. 50. 40. 30. 4812 1 1 1 1 27ll 38 1 2 1 1 1 ""1711 2fU2 ^ 1 l 1 1 1 7'1 ie',2 1 1 1 l 1 1 2nd 3rd 4th 5th 6th 7th Diagram III. — Progress of 2,654 pupils in quality of silent reading 8th answered correctly a grade of 10 points was given. The average of the reproduction grade and the grade received for correct answers to questions was found. This average grade formed the quality score in silent reading. The average quality score was then deter- mined. The progress of 2,654 pupils in quality of silent reading is represented in Diagram III. This diagram contains readjustments similar to those described in connection with Diagram II and should be similarly interpreted. The average quality mark for each of the grades is as follows: "Tiny Tad" Second Third "The Grasshoppers" Fourth Fifth * Sixth 32 37 29 32 39 "Ancient Ships" Seventh Eighth 22 27 DESCRIPTION OF THE READING TESTS 55 Revised score sheet. — The score sheet for both the oral-reading test and the silent-reading tests is presented on the following page in the form in which it is now being used. The directions for giving both tests and for scoring the results as described on the back of the score sheet are given on pp. 55-59). DIRECTIONS FOR GIVING THE ORAL-READING TESTS 1. The pupils should be tested individually in a quiet place where they will be free from distraction and where the remainder of the pupils to be tested will not hear the reading. 2. When everything is in readiness to begin, hand the pupil a copy of the stand- ardized paragraphs and give the following directions: "I should like for you to read some of these paragraphs for me. Begin with the first paragraph when I say ' Begin.' Stop at the end of each paragraph until I say ' Next.' If you should find some hard words, read them as best you can without help and continue reading." Pupils above the fourth grade should begin with paragraph 4. If two or more errors are made in this paragraph, ask the pupil to read the preceding paragraphs also. In case pupils in the first two grades hesitate several seconds on a difficult word, pronounce it for the pupil and mark it as mispronounced. 3. While the pupil is reading, record two sets of facts in regard to the reading: the time required to read each paragraph and the errors made. a) The time record is secured by noting the exact second at which the pupil begins reading a paragraph and the time when he completes it. The number of seconds required to read the paragraph should be recorded in the margin to the right of the paragraph. b) In order to illustrate clearly the character of errors and the method of record- ing them, the following paragraph is inserted: The sun pierced into my large windows. It was the opening of October, tie** /->. - and the A sky was^ofja dazzling blue. I looked out of my window Cand) down the street. The white houseQof the long, slight street were(^nost painful to the eyes. The clear atmosphere allowed full play to jhg sun^sjprightness. If a word is wholly mispronounced, underline it as in the case of "atmosphere." If a portion of a word is mispronounced, mark appropriately as indicated above: "pierced" pronounced in two syllables, sounding long a in "dazzling," omitting the * in "houses" or the al from "almost," or the r in "straight." Omitted words are marked as in the case of "of" and "and"; substitutions as in the case of "many" for "my"; insertions as in the case of "clear"; and repetitions as in the case of "to the sun's." Two or more words should be repeated to count as a repetition. It is very difficult to record the exact nature of each error. Do this as nearly as you can. In all cases where you are unable to define clearly the specific character of the error, underline the word or portion of the word mispronounced. Be sure you put down a mark for each error. In case you are not sure that an error was made, 56 STUDIES OF ELEMENTARY -SCHOOL READING o i—i 3 S -^ o a H w H CO W o o o CO 1 o w « g 3 < w g w d en >» 05 73 g (gen Is O"" S.2 4> P* CO O PS o CJ w PS » g w OS o H bo nj Ph 72 O R *H P-lC/3 M H o Oi 00 t^ VO « >*■ CO 01 » "3, 3 Ph *c3 a 55 60 < in 6 55 H fN CO >+ u-> o w 00 o o H M CN CO CD M o CJ CO "c3 o H DESCRIPTION OF THE READING TESTS 57 give the pupil the benefit of the doubt. If the pupil has a slight foreign accent, dis- tinguish carefully between this difficulty and real errors. 4. Each pupil should be allowed to continue reading until he makes seven errors in each of two paragraphs. DIRECTIONS FOR SCORING THE RESULTS i. Score the results for each paragraph through the use of the following table. The numbers in the left-hand column refer to the number of seconds required to read a paragraph. The numbers in the horizontal line at the top of the table refer to the number of errors made in reading. The numbers in the horizontal line to the right of 40 mean that if a paragraph is read in 40 or more seconds with no errors a credit of 4 is given; with 1 error, a credit of 4; with 2 errors, a credit of 3; with 3 errors, a credit of 2, etc. Seconds 40 or more 30-39 25-29 20-24 19 or less. . Errors 7 or More To find the score for a given paragraph, note the time required to read it and the number of errors made. For illustration, paragraph 1 may be read by pupil A in 34 seconds with 3 errors. In the left-hand column of the table find the time unit which corresponds to 34 seconds. Evidently it is the time unit 30-39. Follow the horizontal line of numbers to the right of 30-39 to the column which represents 3 errors. The score indicated there is 2. Enter this score on the score sheet in the column for paragraph 1, opposite the reader's name. The score for each paragraph should be determined and entered in the same way. Make no entry on the score sheet if the score is o. 2. Proceed as follows to find the average class score: a) Find the sum of the scores for each paragraph separately and enter each total score at the foot of the appropriate column on the score sheet. b) Enter the total score for each paragraph in the column under "Score," in the table on p. 58. c) The value or credit given for the successful reading of paragraph 1 varies with the grade. These values are given in the column to the right of the table. Enter the appropriate value for paragraph 1 in the blank space in the column under "Value." Thus, the appropriate value for paragraph i, for the third grade, is 30. The values for all other paragraphs remain the same for all grades. d) Multiply the score for each paragraph by its value and enter the result in the column under "Product." 58 STUDIES OF ELEMENTARY-SCHOOL READING e) Find the sum of the products and divide by 4 times the number of pupils in the class. The result is the average class score. 3. Individual scores may be found as follows: a) Do as directed in b), c), and d) in the directions for finding the average class score. b) Divide the sum of the products by 4. The result is the individual score. c) The average class score may be found by finding the average of the individual scores. Paragraph Score X Value Product I 2 5 VALUE FOR PARAGRAPH I % 5 4 5 Grade 1 55 c 5 II 35 III 30 6 5 7 5 8 5 IV 25 Q 5 V 20 IO 5 VI 15 vn 10 II 10 12 5 VIII 5 Average class score . . . DIRECTIONS FOR GIVING THE SILENT-READING TESTS 1. Grades II and III are tested on "Tiny Tad;" Grades FV, V, and VI on "The Grasshoppers;" Grades VII and VIII on "Ancient Ships." 2. The teacher then hands the pupil a card with the selection appropriate for his grade printed on it, with these directions: "Read the story on this card silently. Read the story from beginning to end without stopping or repeating any of it. Read the story rapidly but carefully. Do not stop reading to ask about difficult words; read such as best you can. Be prepared to tell the story or to answer any question about it when you are through. Do you understand?" Make the directions clear to the pupil before beginning the reading. 3. Secure the time record as follows: When the pupil shifts from the bottom of the first column to the top of the middle column, note the position of the second-hand. When he shifts from the bottom of the middle column to the top of the third column, note the second-hand again. Record the number of seconds required to read the middle column. Be sure that your record indicates the time for reading the middle column only. If possible, use a stop watch. 4. If the pupil has been reading "Tiny Tad," ask him to tell you the story of Tiny Tad as well as he can. In the meantime write as rapidly as you can just what the pupil says. Ask the pupil to tell a sentence at a time if he talks too rapidly. When he discontinues speaking, ask him if he remembers anything else. Following this, ask him the questions and record his answers. DESCRIPTION OF THE READING TESTS 59 5. If the pupil has been reading "The Grasshoppers" or "Ancient Ships," hand him the report blank to fill out. Ask him to write the story which he has read as well as he can. Impress upon him the fact that he should not look at the questions on the opposite side of the sheet until he has finished writing his story. DIRECTIONS FOE SCORING THE RESULTS Rate of silent reading — 1. Enter the number of seconds required by each Dupil to read 100 words in the column under "Rate" on the record sheet. The time record for pupils who read "The Grasshoppers" and "Ancient Ships" was based on 200 words. Divide the time record for these pupils by 2 before entering the rate on the record sheet. Drop frac- tions. 2. Find the average number of seconds required by the pupils of a class to read 100 words. 3. Express this average in terms of the number of words read per second. To do this divide 100 by the average number of seconds required by the class to read 100 words. Quality of silent reading — 1. Score the reproductions as follows: check from the pupil's reproduction all wrong statements, all irrelevant statements, and all repetitions. Count the remaining words. Find the percentage that these words are of the total nunber of words in the selection. Enter the result in the column under "Reproduction" on the record sheet. 2. Give a credit of 10 points to each question answered correctly. Enter the total grade for each pupil for questions answered correctly in the column under "Ques- tions." 3. Average the reproduction grade and the grade received for correct answers to questions for the quality score. Drop fractions. Enter the quality score in the appropriate column on the record sheet. 4. Find the average quality score for the class by rinding the sum of the individual quality scores and dividing by the number of pupils in the class. CHAPTER IV VALIDITY OF THE ORAL-READING TEST It is the purpose of this chapter to give a description of the methods used in deriving the series of standardized paragraphs for the oral-reading test, to discuss the validity of the test, and to point out significant limitations. The initial attempt to derive an oral-reading test resulted in the selection of a tentative series of paragraphs. This series passed through two successive modifications before the present test was derived. In certain respects the methods which were employed at the outset were retained unchanged throughout the study. In other respects the methods were materially changed as the study developed. The following description of the derivation of the test presents in detail the methods which were used and the results which were secured in each of the successive studies. DESCRIPTIVE SUMMARY OF THE INITIAL STUDY 1 The purpose of this study was to derive a tentative scale for the measurement of certain phases of oral-reading ability. It seemed best to limit this initial investigation to the measurement of ability along certain very definite lines that are objective and easily measured. The time required to read a paragraph and the number of errors made in reading were adopted as the basis for measure- ment. An analysis of the errors made by pupils led to the following classification of errors: i. Gross errors in pronunciation, i.e., mispronunciations which indicate clearly that the words are beyond the pupil's ability. 2. Minor errors in pronunciation, i.e., mispronunciation of one or two elements of a word. 3. Omission of words. 4. Substitution of one word for another. 5. Insertion of words not included in the text. *A Tentative Scale for the Measurement of Oral Reading Ability, Master's Thesis, Teachers College, Columbia University, 19 14. 60 VALIDITY OF THE ORAL-READING TEST 61 Methods employed in selecting the paragraphs for the test. — As a first step in selecting the passages for the reading test ten sets of readers of recent date were studied. From the various readers for the different grades sixty representative paragraphs were chosen with the following points in mind: (i) that each paragraph contain about 50 words; (2) that each paragraph contain a somewhat com- plete thought-unit; (3) that the thought of the paragraphs be adapted to the interests of children. It was impossible to conform strictly to the last point mentioned in the case of the more difficult paragraphs. Furthermore, in a very large number of cases it was necessary to reconstruct the paragraphs in order to make them conform to the three standards mentioned above. From the collection of sixty paragraphs sixteen were chosen, which in the opinion of the writer represented a series of paragraphs of various degrees of difficulty. These paragraphs were then num- bered without regard to their order and submitted to twenty graduate students with these directions: "Please arrange the fol- lowing paragraphs in the order of their difficulty for oral read- ing with emphasis particularly upon pronunciation. When you have completed the arrangement of the paragraphs, record the order on the accompanying blank by placing the number of the easiest slip opposite First, the number of the next easiest opposite Second, etc." The rankings of the paragraphs by the individual judges appear in Table IX. The numbers in the vertical column to the left of the table are the numbers which were assigned to the specimen paragraphs. The numbers in the horizontal line at the top of the table are the numbers which were assigned to the various judges. The entries in the table are the rankings or order of difficulty of the various paragraphs as determined respectively by each judge, 1 representing the easiest; 16, the hardest. Before proceeding farther it should be stated that rankings by judges should not be given much weight in establishing a scale unless the judges are competent to offer expert opinion concerning the problem at hand and unless a large number of judgments are secured upon which to base conclusions. Since the relative difficulty of the paragraphs can be determined most accurately on the basis of data secured from 62 STUDIES OF ELEMENTARY -SCHOOL READING elementary-school pupils, the opinions of twenty graduate students are used in this study merely as a preliminary guide. TABLE IX Distribution of the Rankings of the Specimen Paragraphs by Each of the Judges Para- graph Judges 1 2 3 4 s 6 7 8 9 10 11 12 13 14 is 16 17 18 19 20 I. . . 4 3 3 3 4 4 4 S s 6 4 12 7 2 2 16 6 7 4 3 2 7 6 10 5 6 6 S 9 6 9 7 13 4 5 3 6 7 3 10 4 3 5 4 8 4 5 S 3 6 3 S 9 16 2 4 6 S 4 4 3 6 4 9 9 11 6 8 12 8 13 8 12 14 8 3 6 5 14 11 6 9 7 5 2 1 2 2 2 2 2 2 2 2 4 3 6 3 4 10 1 1 1 1 6 16 16 16 16 16 14 16 16 16 16 16 16 16 16 16 7 16 16 16 6 7 12 10 9 12 11 9 12 3 9 14 8 9 5 9 14 3 8 13 S 9 8 8 12 12 8 10 11 11 IS 12 11 13 10 12 11 8 IS 12 12 7 12 9 6 7 4 11 3 10 6 14 11 4 2 14 10 8 10 8 3 5 6 5 IO 15 IS IS 13 13 8 13 11 14 15 11 4 IS 10 13 13 IS 11 12 13 ii 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 12 14 13 14 15 15 16 14 10 IS 8 IS 2 13 14 9 4 14 15 14 IS 13 11 14 13 14 14 IS IS 7 13 13 12 7 8 IS IS 11 13 14 IS 14 14 10 5 6 9 9 7 7 8 4 10 3 11 11 7 7 9 9 8 11 10 IS 13 11 5 10 12 3 9 S 13 3 6 S 9 12 11 1 10 10 13 8 16. 3 8 7 7 7 13 10 12 7 7 10 IS 14 13 12 12 5 9 8 11 As an aid in determining the combined opinion of the twenty judges in regard to the rank of each paragraph, Table IX was con- verted into Table X, which shows the number of times each para- graph was given certain rankings by the judges. The numbers in the vertical column to the left of the table refer to the paragraphs. The numbers in the horizontal line at the top of the table refer to the rankings. The entries represent the number of times each paragraph was given certain rankings by the twenty judges. To determine temporarily the position of each paragraph in the series, the opinion of the median judge was taken as a guide. Referring now to Table X, paragraph 1, and adding the integers in the line to the right, one must include four-sixths of the 6 in column 4 before reaching the median. Since the median for paragraph 1 falls under rank 4, paragraph 1 was assigned fourth place in the series. In a similar manner the median was found in the case of each of the other paragraphs and their positions* were determined. It will be noted that in the case of paragraphs 4, 8, 9, 13, 14, 15, and 16 the VALIDITY OF THE ORAL-READING TEST 63 position of the paragraph in the series lies between two rankings rather than at a definite point. The median rank of each paragraph is indicated in the vertical column at the right of Table X. TABLE X Number of Times Each Paragraph Was Given Certain Rankings by the Judges Paragraph Rankings Median Rank 1 2 3 4 s 6 7 8 9 10 11 12 13 14 is 16 I O O s 14 I 2 O I O 9 1 6 1 4 2 3 1 2 2 2 1 2 1 6 2 5 2 2 1 1 1 2 3 5 1 2 2 1 3 1 2 5 3 3 1 1 3 1 1 2 3 O I O I O I I O O O 2 4 5 O O I 4 2 3 2 I O I I 2 I 2 O 2 I 3 6 1 4 2 1 2 1 1 2 3 1 1 3 3 2 2 1 4 2 3 2 3 2 1 1 2 3 7 1 r 2 3 1 1 1 1 6 2 4 3 2 2 1 2 2 1 6 S 1 2 6 6 5 1 1 1 17 1 4 2 6 3 5 4 Between 8 and 9 5 2 6 16 7 9 Between 11 and 12 8 9 Between 6 and 7 10 13 1 14 11 12 13 Between 13 and 14 Between 8 and 9 Between 9 and 10 Between 9 and 10 14 15 16 After the relative position of the paragraphs in the series were determined in this general way, the equality of the steps of differ- ence was next determined through a use of the principles involved in The Equality of Equally Often Noted Differences (Thorndike, 1913, p. 122). The method of procedure was as follows: Paragraph 11, Table X, which holds first place in the series, was compared with paragraph 5, which holds second place in the series, to determine how many judges marked 1 1 as harder than 5. Similar comparisons were made between the second and third paragraphs in the series, the third and fourth, etc. In the case of paragraph 9, which holds sixth place in the series, a comparison was made, not only with paragraph 2, which holds fifth place, but also with both paragraphs 4 and 14, each of which holds seventh place in the series. Other double comparisons were necessary, as will be noted in Table XL Table XI is easily interpreted. For our purpose it is essential to note that paragraph 9 is rated as harder than paragraph 4 by 7 judges and as harder than paragraph 14 by 8 judges. Since the 64 STUDIES OF ELEMENTARY-SCHOOL READING differences between 9 and 4, on the one hand, and between 9 and 14, on the other hand, are about equal, according to the opinion of the twenty judges, either paragraph 4 or paragraph 14 may be dropped. For our purposes 14 was dropped. Similarly, the differences between paragraphs 7 and 15 and between paragraphs 7 and 16 are about equal. Either paragraph 15 or paragraph 16 should have been dropped from the list. Through some oversight para- graph 7 was dropped instead. In all other cases the differences between the successive paragraphs are large enough to justify their TABLE XI Steps of Difference between the Successive Paragraphs of the Scale Number of judges rating 11 harder than 5 is 5; vice versa, 15 5 u u I a 2 a ti 18 1 u u 3 a 8 it a 12 3 u u 2 u 5 tt is 2 tt it 9 a 7 a tt 13 9 a it 4 u 7 a tt 13 9 a tt 14 a 8 ti u 12 4 a u 7 a 7 a tt 13 14 tt tt 7 it 7 tt u 13 7 it tt IS it 11 a tt 9 7 a tt 16 It 11 tt tt 9 IS it a 8 a 7 it u 13 16 it tt 8 a 6 tt tt 14 8 it tt 10 tt 7 u a 13 10 a it 13 u 9 u a 11 13 it a 12 tt 8 tt tt 12 12 it it 6 tt 2 u a 18 retention in the list. As a result of the study of the data of Tables IX, X, and XI, the final order of the paragraphs in the revised list was as follows: The intervening numbers inclosed in parentheses represent the steps of difference: n-(i5)-5-(i8)-i-(i2)-3-(i5)-2- (i3)-9-( I 3)-4-(22)-i5-(o or i)-i6-(i4)-8-(i3)-io-(ii)-i3-(i2)- i2-(i8)-6. Two additional steps were now taken in order to complete the scale. It will be noted that the difference between paragraphs 5 and 1 is quite distinct, since 18 of the 20 judges decided that para- graph 5 was easier than paragraph 1. The difference between VALIDITY OF THE ORAL-READING TEST 65 paragraphs 12 and 6 is similarly distinct. Because of the large "steps of difference" between these respective couplets it seemed best to select a paragraph intermediate in difficulty to fit in between the members of each couplet. With the help of a graduate student the needed paragraphs were chosen. This added two new para- graphs to the list, making sixteen in all. The large difference between paragraphs 4 and 15 was due to the error already men- tioned. It was not considered in this revision of the list because the mistake was not discovered in time. Before the test was printed it was given to forty pupils. An examination of the data suggested two changes. First, it seemed that the paragraph which had been inserted between paragraphs 5 and 1 was easier than paragraph 5, and also that there was not quite as distinct a difference between paragraphs 5 and 1 as the judges had thought. This suggested a reorganization of the first four paragraphs of the series into the following order: paragraph 11, the newly inserted paragraph, paragraph 5, paragraph 1. The data also suggested the fact that the scale did not include as diffi- cult passages as would be necessary to test upper-grade pupils. To remedy this situation, four difficult paragraphs were added to the scale. The paragraphs were now numbered in the order of increas- ing difficulty and printed as they appear below. It was time for winter to come. The little birds had all gone far away. They were afraid of the cold. There was no green grass in the fields, and there were no pretty flowers in the gardens. Many of the trees had dropped all their leaves. Cold winter with its snow and ice was coming soon. Once there lived a king and queen in a large palace. But the king and queen were not happy. There were no little children in the house or garden. One day they found a poor little boy and girl at their door. They took them into the palace and made them their own. The king and queen were then happy. 3 Once a green little leaf was heard to sigh and cry, as leaves often do when a gentle wind blows. "What is the matter, Little Leaf ? " said the twig. The little leaf replied, "The wind just told me that one day it would pull me off and throw me down to the ground to die. That is why I am so sad." 66 STUDIES OF ELEMENTARY-SCHOOL READING 4 Once I went home from the city for a summer's rest. I took my gun for a stroll in the woods where I had shot many squirrels. I put my gun against a tree and lay down upon the leaves. Soon I was fast asleep, dreaming of a group of merry, laughing children running and playing about me on all sides . 5 One of the most interesting birds which ever lived in my bird-room was a blue jay named Jakie. He was full of business from morning till night, scarcely ever still. He had been stolen from a nest long before he could fly, and he was reared in a house long before he had been given to me as a pet. 6 Henry was a busy farmer. His fafmhouse stood on a hillside above the seashore. Along the shore and up the hillside were the houses of his friendly neighbors. Around his house the ground was flat, like the top of a huge step in the hillside. All about him stretched his small, verdant rice fields. 7 It was a glad summer morning. Little birds teetered on the twigs of the trees. They opened their throats and sang as loud as they could. Flowers nodded to each other in the gardens and along the wayside. Butterflies went flitting about gayly, the morning air was fresh and sweet, and all was gladness. I remained there nearly two hours, I dare say. Once I opened the yard gate and looked into the empty street. The sand, the seaweeds, and the flakes of foam were driving by, and I was obliged to call for assistance before I could shut the gate again and make it securely fast against the strong wind. 9 The part of farming enjoyed most by a boy is the making of maple sugar. It is better than blackberrying and almost as good as fishing. One reason he likes this work is that someone else does most of it. It is a sort of work in which he can appear to be very industrious, and yet do but little. Rip would carry a fowling piece on his shoulder for hours together. He would trudge through the woods and swamps, up hill and down dale, to shoot a few squirrels or pigeons. He would never refuse to assist a neighbor in any way. Even the women of the village often employed him to run their little errands. VALIDITY OF THE ORAL-READING TEST 67 As far as the eye could reach the sea was of a deep blue color in every direction. The waves were running high, and were fresh and sparkling in the sunlight. In the midst lay an immense iceberg. Its cavities and valleys were thrown into deep shades. Its points and towers glittered brightly in the sun. The sun pierced into my large windows. It was the opening of October, and the sky was of a dazzling blue. I looked out of my window and down the street. The white houses of the long, straight streets were almost painful to the eyes. The clear atmosphere allowed full play to the sun's brightness. 13 The success of Greeley's paper was immediate and great. It grew a little faster than the machinery for producing it could be provided. The success of the paper was due to the fact that the editor's original idea was carried out. He aimed to produce a paper which was morally helpful to the public. 14 It was one of those wonderful evenings such as are found only in this magnificent region. The sun had sunk behind the mountains, but it was still light. The twilight glow embraced a third of the sky, and against its brilliancy stood the dull white masses of the mountains in evident contrast. 15 The crown and glory of a useful life is character. It is the noblest posses- sion of man, constituting a rank in itself, an estate in the general good-will, dignifying every station, and exalting every position in society. It exercises a greater power than wealth, and is a valuable means of securing honor. 16 George Washington was in every sense of the word a wise, good, and great man. But his temper was naturally irritable and high-toned. Through reflec- tion and resolution he had obtained a firm and habitual ascendancy over it. If, however, it broke loose its bonds, he was most tremendous in his wrath. 17 He was six feet tall, and his body was well-proportioned. His complexion inclined to the florid; his eyes were blue and remarkably far apart. A pro- fusion of hair covered the forehead. He was scrupulously neat in his appear- ance, and, although he habitually left his tent at an early hour, he was well dressed. 68 STUDIES OF ELEMENTARY-SCHOOL READING Responding to the impulse of habit, Josephus spoke and the others listened attentively, but in grim and contemptuous silence. He spoke for a long time, continuously, persistently, and ingratiatingly. Finally exhausted through lack of nourishment, he hesitated. As always happens in that contingency, he was lost. 19 The hypotheses concerning physical phenomena formulated by the early philosophers proved to be inconsistent and, in general, not universally appli- cable. Before relatively accurate principles could be established, physicists, mathematicians, and statisticians had to combine forces and work arduously. "Read the following sentences correctly: Sophistry is fallacious reason- ing. They resuscitated him. Verbiage is wordiness. Equanimity is evenness of mind. He has a pertinacious, obstinate disposition. There was subtlety and poignancy in his remarks. A hypocritical and pharisaical nature is usually cynical." To whom the test was given. — The test was given to 565 pupils from the third to the eighth grade inclusive. These pupils repre- sented four schools, three in the city of New York and one in a small city in central Illinois. Two of the schools of New York are located in foreign districts of the city. The third New York school represents a more truly American population, economically inde- pendent. The Illinois school represents an American population of average economic rank. The pupils were about equally divided between native American and foreign-born children, and represent practically every economic level. Scoring the results. — The method of securing the data and scoring the results was the same as that described in chap, hi of this report. One exception should be mentioned. In this initial study repeti- tions were not recorded as a type of error. In a later section of this chapter the validity of the standards for scoring will be discussed. The records of all pupils were scored according to each of the four standards. Certain of the results are given in Table XII. The table contains the results for the sixth grade by each of the four standards, the aggregate results for the sixth grade by all standards, and the aggregate results for all grades by all the VALIDITY OF THE ORAL-READING TEST 69 ►J J < e S 9 < ae Men as IMS) W H o« iO CI lO cOO O « M fOM O.M00 WjOO Oi« )M O^O W)(Oro« mm OO <00 Oi O W© *■* I I I I I I I I I I I ^•NM iflOiM>0 O>00 O O rf r^.00 tHO^O *00 000 r^ r*o O 10 »o «n vj >o >o 't ^t ro w w m O O r-* *^> *>*> *t fO <0 < O o> r* *o *o ro IO OO «*» m l>. m < l l l I l l I l l I l I I 1 I+++++ IVOOO ^OCO MOO ^tl 1 ^fOO « NNOi^N com rod M « 10 10 O 0» 000 00 00 00 t-*00 t»» t*» r* t*« r-»o 10 -^- ^t < H («)lfl«0 fOOO H NOO tflNM M C 0> -■■ tO O fON'tNH ^- ^to O m < ^■a^M t^oo o 1 1 1 1 1 1+1 1 1 1 1 1 10O * t-*. O *■*■■ m 00 m moo 00 *o O to *fr o 00 00 00 o mo o mo moo o uTtMOO -Omom too ooiO>o« 000 cooooooooooooo t^mrj-mfo 'twooo O o\ m o 00 00 wj'+'t'tO'N 000 00 O m m O O 00 00 000 *--oo t-^ r^ m « o m^o • O moooooo O &■ chc> r*. w m c « t-co to • wmo^oO^OdOO 00 o m t^ to tooo OOOmmoOHcO O O 00 OoO moo u-> . O O COCOO-COOO O>00 CO M I I I I I I I I ++ hnQm fOTj-iri tOO OO O r^.o r» moo f* ci O O fO 't < . M C OO OOO N« tOO OO O o o coco 000 mo i6i The second purpose of the preliminary test was to determine which of the selections would form a series increasing in difficulty by appropriate steps of difference. The data secured in this study did not prove entirely adequate for this purpose, because the num- ber of pupils who read the same two successive paragraphs was small in each case. It was decided, therefore, to choose those selections which had proved most successful in a general way in the preliminary study. Upon this basis "Tiny Tad," "The Grasshoppers," and "Ancient Ships" were chosen. Table XXVII shows that these selections increased in difficulty quite regularly both as to speed and as to quality for the successive grades. The preliminary study offered a number of valuable suggestions concerning the more effective organization of silent-reading tests. Certain radical changes in the structure of the subject-matter and VALIDITY OF THE SILENT-READING TESTS 103 in the character of the questions were shown to be desirable. Practically all teachers affirmed that a test which requires second- and third-grade pupils to write their answers to the questions penalizes these pupils unduly, because writing in itself presents many difficulties at this stage of the development of pupils. Many teachers questioned the reliability of the rate when it was indicated by the pupil's own mark. Concrete cases were cited in justification of this criticism. Several teachers tested the pupils on their com- prehension of what was read in other ways than by means of ques- tions. They concluded from the results that questions alone do not afford a sufficiently broad basis for testing comprehension. Finally the suggestion was offered that it frequently required several seconds for pupils to adjust themselves to the conditions and subject-matter of the test, and that pupils should be allowed to read several seconds before the experimenter begins the time record. As a result of this preliminary study and the suggestions which grew out of it, the silent-reading tests were organized as presented in chap. iii. For the purpose of this thesis it was very desirable to devise a test which would be as accurate as possible. Wherever questions arose, ease of giving the tests was usually sacrificed for increased accuracy in results. Different characteristics of the test will now be taken up in turn and discussed from the standpoint of validity. CRITICAL DISCUSSION OF THE ADOPTED METHODS OF TESTING SILENT READING Form in which the test was printed. — To obviate the difficulty presented by the fact that the pupil's own record of his achieve- ment is sometimes inaccurate, the revised test was given indi- vidually and the rate was recorded by the teacher. This method has the negative advantage that it does not interrupt the pupil's reading at any point. The positive advantages of the method are that it enables the teacher to make a personal study of the pupil while he is reading, and it imposes the responsibility of accuracy upon the teacher. Pupils have many habits connected with their reading, such as excessive lip movement, pointing with finger, and 104 STUDIES OF ELEMENTARY-SCHOOL READING poor position of the book, which are positive handicaps to effective reading. Furthermore, some pupils are unable to shift the eye accurately from the end of one line to the beginning of the next. Personal study and observation of the reader frequently enable the teacher to determine the causes of inefficiency. Since it is more important to know why a pupil fails than it is to know simply that he is failing, the advantages of an individual test more than out- weigh the disadvantages of greater time and labor involved, par- ticularly for the purpose of the present study. The difficulties which arose from the fact that pupils vary in the time at which they begin and discontinue their reading and in the readiness with which they adapt themselves to the subject-matter which is to be read and to the conditions of the test were obviated in the following ways : The subject-matter of each test was printed on a cardboard in three columns. The middle column contained ioo words in the case of "Tiny Tad" and 200 words each in the case of "The Grasshoppers" and "Ancient Ships." The short column to the left enabled the reader to assume his natural rate of reading before the time record began. Variations in rate due to the fact that pupils begin and discontinue their reading at slightly different time periods were obviated by the fact that each pupil's record was secured while he was reading at his normal rate in the middle of the selection. The fact that the reader had to shift his eyes from the bottom of the first column to the top of the second and again from the bottom of the second column to the top of the third enabled the investigator to note accurately the time at which the reading of the second column began and ended. Length of the test. — Several points were considered in determining the length of the reading tests. In the first place, practical con- siderations require that the shortest test be used which will secure accurate results. Concerning the relation of the length of the test to the rate of reading, Starch (Journal of Educational Psychology, VI, 8) shows that the number of words read per second is practically the same whether a person reads for 30 or for 60 seconds. He therefore concluded that a test of 30 seconds was sufficiently long to insure accuracy. Since many pupils read the selections involved in this test in less than 30 seconds, the following study was made : VALIDITY OF THE SILENT-READING TESTS 105 A group of eighth-grade pupils were tested first upon " Ancient Ships" and then upon a selection the middle column of which con- tained 500 words. The results of this test showed that the average rate of reading was the same whether the middle column contained 200 words or 500 words. It was concluded, therefore, that the tests as devised were sufficiently long to serve as accurate measures of rate. Tests conducted by Miss Katherine McLaughlin in the Ele- mentary School of the University of Chicago show that fifth-grade pupils reproduce and answer questions concerning the latter por- tions of a selection less well than concerning earlier portions if the selection is over 300 words in length. This proved to be true if either the earlier or the latter portions were reproduced first. She concluded that fatigue or memory span was a real factor in tests which involve long selections. For the purpose of eliminating as many variable elements as possible it was decided to limit the silent-reading tests to less than 300 words. The selection for the second and third grades was made 100 words shorter than the selections for the upper grades. This reduction was made because pupils of the lower grades read much more slowly than pupils of the upper grades. Reliability of a single record for rate. — To determine the reliability of a single record for rate, the following investigation was made: A class of 19 pupils was tested on each of two selections, the one somewhat more difficult than the other. Table XXVIII gives the results for each selection in terms of the number of seconds required by each pupil to read 200 words (see p. 106). The correlation between the two rankings, as computed by Spearman's footrule, is +0.88. This indicates a high degree of accuracy in a single record for rate. Method of determining rate. — In order that the average achieve- ment of groups of individuals as to rate of silent reading might be represented numerically, the harmonic mean of the rates (the arith- metical mean of absolute times) was adopted. This average was then expressed in terms of the number of words read per second. Comprehension based on reproduction and answers to questions. — When a pupil tries to understand what he reads, a number of com- plex processes are involved. The entire meaning which he secures io6 STUDIES OF ELEMENTARY-SCHOOL READING is the result of this total mental activity. Meaning may take various forms, such as general impressions, feelings of assent or disapproval, a mental record of the most important points, a mental record of practically all the details, a recognition of the relationship of the facts expressed to other known facts, etc. It would be impracticable in an investigation of this type, even if it were pos- sible, to test meaning in all its forms, on account of the amount of TABLE XXVIII Influence of Different Selections on Rank in Silent Reading Selection A Selection B Pupil Time in Seconds Rank Time in Seconds Rank i 17 20 20 22 22 22 23 29 3° 3° 3° 31 32 32 34 35 38 40 40 I 2 3 4 5 6 7 8 9 10 11 12 13 14 IS 16 17 18 19 23 27 31 26 28 29 29 32 31 33 43 43 47 3° 45 40 60 57 59 I 2 3 8 ■z 4 2 5 4 5 6 6 7 8 10 9 9 11 IO ii 13 14 16 12 13 14 7 15 12 15 16 17 19 17 18 18 19 time which would be required. For the purpose of this study it seemed best to test by at least two methods. Reproduction and answers to questions were the means adopted because they repre- sent two of the most frequent ways in which comprehension is tested in our schools today. That both methods form a better test than either one alone is shown by the fact that pupils vary widely in their ability to reproduce and to answer questions. The records in Table XXIX show that some pupils do relatively well by one method and relatively poorly by the other. ■ VALIDITY OF THE SILENT-READING TESTS 107 Oral reproductions below the fourth grade. — The fact has been pointed out that second- and third-grade pupils are greatly handi- capped by the requirement that they write their own reproductions. Tests were given in all grades from the second to the sixth inclusive TABLE XXLX Relative Achievement of Pupils in Reproductions v and Answers to Questions Grade Pupil Reproduction Score Question Score f a 14 30 LIB \ b 12 30 I c 6 30 a 30 60 b 25 60 c 22 60 VIA \ d 17 60 e IO 60 f IS 80 8 27 60 h 28 40 to determine the relative efficiency with which pupils report orally and write their reproductions. The pupils were first tested on their ability to reproduce a selection orally. Upon the basis of these results the class was divided into two groups of equal ability. TABLE XXX Comparison of the Percentage Reproduced Orally with Percentage Reproduced in Writing Selection Grade Written Reproduction Oral Reproduction Tiny Tad Tiny Tad The Grasshoppers . The Grasshoppers . The Grasshoppers . II III IV V VI 5-0 14.O 14-3 17.0 20.3 20.O 25.0 15-8 18.0 20.5 One section of each grade then read "Tiny Tad" or "The Grass- hoppers" and reproduced the selection in writing. The second section of each grade gave the reproduction orally. Table XXX gives the average reproduction scores for each section of each grade. 108 STUDIES OF ELEMENTARY-SCHOOL READING The table shows that pupils in the second and third grades are dis- tinctly handicapped by the requirement that they write their repro- ductions. Even in the fourth, fifth, and sixth grades those who report their reproductions orally do slightly better than those who write them. The difference is so small, however, that it need not be considered seriously. In the light of this evidence the method was adopted of having second- and third-grade pupils report their reproductions orally and of having the pupils of all other grades write. Method of scoring reproductions. — The method used in this investigation for scoring reproductions was as follows : All errone- ous statements, all repetitions, and all irrelevant statements were checked from the pupils' reproductions and the remaining words were counted. The ratio of these words to the total number of words read was taken as the reproduction score. This method is very similar to the "word-counting method" adopted by Starch (Journal of Educational Psychology, VI, 7) . The only difference is that Starch takes the total number of words reproduced as the index of comprehension, while the method followed here takes the percentage of words correctly reproduced as the index of com- prehension. A sample test follows to illustrate the method of scoring. It is a test of a second-grade girl. She wrote 69 words, 13 of which were checked from her reproduction. Since she read 174 words and reproduced 56 of them correctly, her reproduction scores 32 per cent. Tiny Tad was a queer little fellow. He was nearly black. He said, "When I grow up and get my legs I am going to hop around to the orchard. I wish I had a brother to hop with me." One morning a little toad scrambled out on the bank. It was Tiny Tad. He had grown his hind legs, and woo now hopping around aa any oth e r toad would . For the purpose of checking the results of this method of scoring with the "group-idea method" which has been widely used, the selections were analyzed according to the suggestions of Judd (Elementary School Teacher, XIV, 371). The significant feature of the method is that the unit of measurement is a group of ideas in significant relationship. After the selections had been analyzed VALIDITY OF THE SILENT-READING TESTS 109 into group-ideas according to the judgment of the writer, the groups of ideas selected were compared with the grouping which appeared most frequently in the pupils' reproductions. Whenever there was doubt concerning the best grouping or whenever the grouping selected differed materially from that given by the pupils, the practice of the majority of the pupils was accepted as the standard. The final analysis of "Tiny Tad" was as follows: Tiny Tad Tiny Tad was a queer little fellow | with only two legs | and a short ta.il.) He was nearly black, too, | and much smaller than most tadpoles | in the big pond. I He could hardly wait | for his front legs to grow. | "When I have them all," | he said, | "I'll leave this dirty water | and go up into the orchard. | What fun it will be | to hop and hop and hop. | If only I had a little brother | to hop with me, | I should be so happy." | It wasn't long | before his legs began to grow. | He moved about | and kicked around | until his legs were quite strong. | "I am going out on the bank | to see if I can hop," | he said | one night | when he was just six weeks old. I The sun was hardly up | the next morning | when a little toad jumped out of the water | and hopped up on the bank. | He was very small, | but none too small for the little legs | that wabbled under him. | It was Tiny, | the young toad. | According to this analysis, each sentence is divided into several group-ideas, each of which when accurately reproduced is given full credit. If a group-idea is partially reproduced, it is given half- credit. If a totally new or wrong idea is introduced, it is checked out. If a group-idea in the reproduction represents an accurate combination of portions of the ideas in two or more groups, it is given full credit for one, one and one-half, or two group-ideas of the original text, according to its content and quality. The follow- ing illustration will make clear this method of scoring. The num- bers to the left of the lines represent the amount of credit given for each group of ideas. Reproduction by a Third-Grade Pupil 1 "Tiny Tad was a queer little fellow." Full credit. .1 "He was nearly black." Full credit. no STUDIES OF ELEMENTARY-SCHOOL READING z "He said." Full credit. I "When I ( grow up and ) get my legs." Equivalent to "when I have them all." 1 "I am going to hop ( around ) in the orchard." Two groups incompletely reproduced. Given credit for one. 2 "I wish I had a brother to hop with me." Full credit for two group ideas. | "One morning." Not so specific as "the next morning." i "A little toad scrambled out of the pond." Portions of two groups. Given credit for one. i "It was Tiny Tad." Full credit. | "He had grown his hind legs." A true statement but given out of connection. "And was now hopping around as any toad would." A true statement, but none similar to it is made in the original, and hence it is not counted. According to this analysis, 10 of the 36 group-ideas are repro- duced. Reduced to a percentage basis, the reproduction score is 28 per cent. According to the word-counting method, 56 out of 174 words were reproduced correctly. Expressed on a percentage basis, the reproduction score is 32 per cent. According to this illustration, the group-idea method which endeavors to evaluate more accurately the various phases of the reproduction is a more severe method of scoring than is the word-counting method. In order that the similarity of results secured by the two methods might be determined, the reproductions of 220 pupils in six grades of the Elementary School of the University of Chicago were scored by both methods. The results are given in Table XXXI. The table shows that the average scores by the word-counting method are higher throughout the grades. The differences between the two scores are remarkably uniform. The reproductions were scored a second time by the two methods. The class averages were practically the same in each case as those which were secured by the first scoring. The individual scores varied very little for the two scorings by the word-counting method. Some variations appeared in the scores by the group-idea method. It therefore VALIDITY OF THE SILENT-READING TESTS in Grade No. of Pupils Counting Words Group-Idea Differ IIB 12 8.1 75 IIA 15 25.2 22.0 3- IIIB 14 29.O 26.3 2. IIIA !5 29.O 26.0 3- IVB 28 3i-7 28.6 3- IVA 20 35-3 320 3- IVA 20 20.0 16. 5 3- VB 16 23.0 18.5 4- VA 21 25.0 21-5 3- VIA 25 24.6 20.6 4- VIIA 17 28.0 24 .0 4 VIIA 17 25.0 22.0 3 appeared that the word-counting method was the more constant measure as well as the more rapid measure. TABLE XXXI Comparison of Reproduction Grades When Scored by the "Word-Counting" Method and by the "Group-Idea" Method Selection Tiny Tad Tiny Tad Tiny Tad Tiny Tad Tiny Tad Tiny Tad IVA 20 35-3 32.0 3.3 The Grasshoppers IVA 20 20.0 16.5 3.5 The Grasshoppers VB 16 23.0 18.5 4.5 The Grasshoppers VA 21 25.0 21.5 3.5 The Grasshoppers . . . The Grasshoppers . . . Ancient Ships Amount of credit given for correct answers to questions. — A credit of 10 points was given for each question answered correctly. Theoretically this method of grading would be accurate provided the questions were of equal difficulty. Now, in fact, the questions are not of equal difficulty if we take the number of times each question was answered correctly as an index of difficulty. How- ever, the relative difficulty of the questions differed for different schools. The experiences of individuals and classes differ widely, and it is legitimate to expect that some questions will be less diffi- cult for some than for others. Hence the course of procedure which was adopted for this study was to secure a list of questions so diffi- cult that very few pupils of a class would be able to answer all the questions and at the same time easy enough so that no pupil would fail entirely unless he were an exceptional case. After a long period of experimentation, the questions used in this study were adopted. The number of times that each question was answered correctly by the first one hundred pupils who took each test after the final lists of questions had been chosen is shown in Table XXXII. The numbers show that the questions which were selected fulfil the conditions mentioned above. After the desired degree of uni- formity in the difficulty of the questions had been secured, it was 112 STUDIES OF ELEMENTARY-SCHOOL READING arbitrarily decided to give a credit of 10 points for each question correctly answered. TABLE XXXII Relative Difficulty of the Questions for Each Selection Question Tiny Tad The Grasshoppers Ancient Ships I 74 74 55 47 8o 4i 43 3° 53 20 69 51 63 29 5° 48 52 60 39 58 SO 70 84 2 3 4 75 60 5 6 30 48 56 40 42 7 8 9 IO VALIDITY OF THE STANDARDS OF ACHIEVEMENT IN SPEED AND QUALITY OF SILENT READING In order to determine the steps of difference in difficulty between the respective selections and to derive standards of achievement for each grade, 2,654 tests were given in 13 cities of Iowa, Minne- sota, Tennessee, and Illinois. In general, the same instructions were used in giving these tests as are outlined in chap. iii. One important addition should be mentioned. To determine the rela- tive difficulty of the various selections, it was necessary to have certain grades read two selections. Therefore fourth-grade pupils were asked to read both "Tiny Tad" and "The Grasshoppers," and sixth-grade pupils were asked to read both "The Grasshoppers" and "Ancient Ships." After all the tests had been given, a record was made of the number of seconds required by each pupil to read 100 words. These ' rates were tabulated in order from the most rapid to the slowest. For the purpose of this report the records have been grouped as follows: Those pupils who required from o to 4 seconds, from 5 to 9 seconds, from 10 to 14 seconds, from 15 to 19 seconds, etc., were grouped together. The results for 2,654 pupils appear in Table XXXIII. VALIDITY OF THE SILENT-READING TESTS "3 TABLE XXXIII Distribution of 2,654 Pupils with Reference to the Time Required to Read 100 Words Silently No. of Seconds No. in Grade II No. in Grade III No. in Grade IV No. in Grade IV No. in Grade V No. in Grade VI No. in Grade VI No. in Grade VII No. in Grade VIII o- 4. 5- 9- 10- 14. 15- 19- 20- 24. 25- 29. 3°- 34- 35- 39- 40- 44. 45- 49- So- 54- 55- 59- 60- 64. 65- 69. 70- 74. 75- 79- 80- 84. 85- 89. 90- 94. 95- 99- 100-104. 105-109. 110-114. 115-119. 120-124. 125-129. I3°- I 34- I35-I39- 140-144 . 145-149. i50~ I 54- 155-159- 160-164. 165-169. 170-174. 180-184. 185-189. 190-194. 195-199. 200-204 . 205-209 . 225-229. 285-289. 4 4 5 14 17 17 21 32 34 38 25 24 23 23 15 17 6 10 7 9 5 10 6 3 Total.... Median time Average time Words per second . 389 66 66.6 i-5 7 25 27 57 44 39 34 46 12 34 17 21 10 7 6 7 3 5 1 3 1 1 6 18 37 42 72 5i 46 24 22 15 10 12 4 3 24 19 46 36 38 45 35 27 27 13 3 16 5 7 5 7 3 32 43 79 64 52 35 32 25 25 73 64 63 49 42 20 16 5 6 4 4 18 4i 54 66 54 50 3i 18 15 6 7 3 1 1 2 16 57 86 74 52 34 3° 23 16 5 2 4 3 1 417 45 43-4 2-3 373 35 33-4 2.99 373 45 45-5 422 39 38.9 2-57 377 37 35-8 2.79 377 40 39-2 2-55 2 11 21 38 60 44 40 19 16 4 4 6 405 37 37-2 2.69 271 35 34-8 2.87 ii4 STUDIES OF ELEMENTARY-SCHOOL READING From the facts presented in Table XXXIII it has been possible to plot a curve of progress for rate of silent reading. Diagram VIII represents the progress of 2,654 pupils in silent reading and may be used as a standard for comparison. The diagram was constructed as follows : The vertical lines in the table represent the grades from the second to the eighth inclusive. The numbers 1 . 00, 2 . 00, 3 . 00, and 4.00 to the left of the diagram designate the rates at which second- and third-grade pupils read "Tiny Tad"; the numbers 1 . 21, 2.21, and 3.21, between the third and fourth grades, designate 4.00 1 1 1 • 3121 1 1 t 1 2i97 3.00 ■ 1 1 1 1 1 2J21 / x^^ - _^~~ 1 _ — r~ 1 1 li97 2.00 1 yf If ^r 1 l{21 y 1 1 1 1 1 i97 1.00 1 ■ 1 1 1 1 1 1 1 1 2nd 3rd 4th 5th 6 th 7 th Diagram VIII. — Progress of 2,654 pupils in rate of silent reading 8th the rates at which fourth-, fifth-, and sixth-grade pupils read "The Grasshoppers;" and the numbers .97, 1.97, and 2.97, between the sixth and seventh grades, designate the rates at which seven th- and eighth-grade pupils read "Ancient Ships." These readjust- ments in the diagram were made as follows: Fourth-grade pupils read "Tiny Tad" at a rate of 2.99 words per second and "The Grasshoppers" at a rate of 2. 20 words per second. The numbers designating the rates for fourth-, fifth-, and sixth-grade pupils were shifted along the vertical axis until a rate of 2 . 20 words per second for "The Grasshoppers" was equivalent to a rate of 2.99 words per second for "Tiny Tad." Similarly, the numbers designating VALIDITY OF THE SILENT-READING TESTS 115 the rate for seventh- and eighth-grade pupils were so arranged along the vertical axis that a rate of 2.55 words per second for sixth-grade pupils reading "Ancient Ships" was equivalent to a rate of 2 . 79 words per second for "The Grasshoppers." Upon the basis of this organization the curve of progress was drawn as repre- sented in the diagram. In order to determine the extent to which this assumed curve of progress represents the real curve of progress through the grades, 200 pupils from the fourth to the eighth grades inclusive were tested on "Tiny Tad." The results are represented by the dotted line in the diagram. The close correspondence between the assumed curve and the curve for the pupils involved in the check test justifies the belief that the assumed curve repre- sents quite accurately the increase in rate of silent reading through the grades. The reproductions and the answers to questions were scored according to the method described on p. 54. The average of these two scores for each pupil was adopted as the comprehension score. The comprehension scores were then arranged in order from the lowest to the highest. They are presented in summary form in Table XXXIV. The progress in quality of silent reading for 2,654 pupils is represented in Diagram IX and may be used as a basis for com- parison. This diagram was constructed and should be interpreted in the same way as Diagram VIII. To determine the extent to which this assumed curve represents the real curve of progress in quality of silent reading, the reproduc- tion and answers to questions were scored for 200 pupils in the grades from the fourth to the eighth inclusive. These results are represented by the dotted line in Diagram IX. Again the very close correspondence between the assumed curve and the real curve would indicate that the former represents with a fair degree of accuracy the curve of progress in quality of silent reading. Limitations of the silent-reading tests. — The silent-reading tests measure three things: namely, rate of reading a given selection, ability to reproduce what is read, and ability to answer specific questions concerning the subject-matter. To test comprehension in a thoroughgoing way, a number of selections should be used n6 STUDIES OF ELEMENTARY-SCHOOL READING rather than one, and a larger number of tests of what one gets out of his reading should be included. The selection entitled "Ancient Ships" is relatively more diffi- cult for pupils in the seventh and eighth grades than are "The Grasshoppers" and "Tiny Tad" for pupils in the intermediate and lower grades, respectively. More satisfactory, but perhaps no more accurate, results might have been secured had the selection for the upper grades been less difficult. TABLE XXXIV Distribution of 2,654 Pupils with Reference to Scores for Quality of Silent Reading Comprehension Score No. in Grade II No. in Grade III No. in Grade IV No. in Grade IV No. in Grade V No. in Grade VI No. in Grade VI No. in Grade VII No. in Grade VIII O- 4.. 5- 9-- 10-14. . 15-19. . 20-24. . 25-29 ■ • 30-34 ■ ■ 35-39- • 40-44 . . 45-49 • • 50-54-. 55-59- • 60-64. . 65-69 . . 70-74.. 75-79 • • 80-84 • • Total... Median . Average 4 13 19 34 4i 60 57 54 42 24 15 6 8 6 3 2 1 389 3i 31-5 5 15 14 33 50 58 65 5o 38 4i 24 14 16 22 27 46 54 60 40 41 30 17 4 3 16 22 30 35 50 49 39 38 39 26 14 10 4 7 24 26 44 40 48 5i 38 45 39 35 16 6 4 6 10 17 29 35 44 44 48 44 32 32 12 10 9 61 66 63 50 32 27 18 13 417 37 37-3 373 40 40.6 373 28 28.8 422 32 31-8 377 39 39-i 17 35 58 67 64 42 40 26 18 ii 377 18 18.0 405 22 21.6 9 25 32 34 24 20 32 24 25 20 9 7 4 2 4 271 26 26.9 The tests are given individually and therefore require consider- able time and effort. The method of scoring results is slow and somewhat laborious. The tests are therefore not so well adapted to practical schoolroom purposes as are many of the tests described in chap. ii. From the point of view of their value as aids in educa- tional diagnosis and supervision these tests are distinctly superior to some which are more easily administered, because they afford to the teacher an excellent opportunity to make a careful study of the difficulties of each pupil during the course of the test. For VALIDITY OF THE SILENT-READING TESTS 117 the purpose of this investigation they have made it possible to secure a body of very valuable data. The method used in this study for scoring reproductions is open to the criticism that unless investigators are carefully trained the scores given by different investigators to the same reproduction will vary. For the purpose of determining the extent to which this criticism is true, ten graduate students were asked to grade each of ten reproductions. The average variation of the individual 60. 1 1 i 1 1 4812 27ll 50. 1 l 1 • 1 38)2 1 1 40. 1 1 28 J 2^-" ico>-tcoO<^OOOOi-i h m c IS -r * C ,A - c ■ r i c- ■ T ■) r cn 4 CA C 1- r- ■ • ■ 134 STUDIES OF ELEMENTARY-SCHOOL READING "slow," and the middle third by the term "medium." In a similar manner the quality records were divided into tertiles, and the thirds were designated, in order, "good," "medium," and "poor." Upon the basis of this classification an individual record must fall into one of the following nine classes: rapid speed and good quality; rapid speed and medium quality; rapid speed and poor quality; medium speed and good quality; medium speed and medium quality; medium speed and poor quality; slow speed and good quality; slow speed and medium quality; slow speed and poor quality. The percentage of each grade which fell in each of the nine classes was calculated. The average percentage of cases for TABLE XXXVIII Average Speed and Quality Scores in Silent Reading of 1,831 Cleveland Pupils and of 2,654 Pupils of Other Cities Grade Rate for Cleveland Rate for Thirteen Cities Quality for Cleveland Quality for Thirteen Cities II I.96 2-57 2.76 2.74 3.00 2.89 3-07 I -5° 2.30 2.20 2-57 2.79 2.69 2.87 25.2 32.I 17-5 22.0 26.1 19.7 24. 2 31-5 37-3 28.8 Ill IV V 31-8 39- 1 21.6 26.9 VI VII VIII all the grades falling in each of the nine classes was then deter- mined. The results are presented in Diagram XIX. This diagram emphasizes the fact that good readers are usually not slow and poor readers are usually not fast. It is evidently not safe to lay down any absolute rule. The average teacher is con- fronted with pupils who might fall in any one or all of the nine classes mentioned. There are good readers who are rapid and there are good readers who are slow. There are rapid readers who retain much of what they read and there are rapid readers who retain little. The points of greatest significance revealed by the table are the facts that high rate and good quality are more com- monly related than slow rate and good quality, and that slow rate and poor quality are more commonly related than rapid rate and poor quality. THE INVESTIGATION OF READING IN A CITY SYSTEM 135 5.00 4.00 3.00 2.00 1.00 1 1 1 1 1 3121 2«97 1 1 i 1 / • / 1 / 2*21 1 _ . 1'.97 s / 1 y f yT ^r 1 S • It 21 197 ' / 1 1 1 1 -21 2nd 3rd 4th 5th 6th 7 th 8 th Diagram XVII. — Average scores in the rate of silent reading of 1,831 Cleveland pupils and of 2,654 pupils of thirteen other cities. Dotted line shows Cleveland scores and solid line those of other cities. — 17^1 •' ^S^ s' y/^ S" ^ ** ** 50 38^2 1 / ,/ /| / | 40 28 J 2 ^ / 7jl 30 s """Tfij^ f 20 2nd* 3rd 4th 5th 6th 7 th 8 th Diagram XVIII. — Average scores in the quality of silent reading of 1,831 Cleve- land pupils and of 2,654 pupils of thirteen other cities. Dotted line shows Cleveland scores and solid line those of other cities. 136 STUDIES OF ELEMENTARY-SCHOOL READING The Cleveland Survey monograph entitled Measuring the Work of the Public Schools reports a similar study in which the records included in this investigation were distributed as follows: In the case of speed and of quality 25 per cent of the records were included in the best group, 50 per cent of the records in the medium group, and 25 per cent of the records in the poorest group. The average percentage of records falling in each of the nine classes was deter- mined as outlined above. The percentage for each class was as follows: rapid speed and good quality, 10; rapid speed and medium 14* 11* e* Rapid Speed and Good Quality Medium Speed and Good Quality Slow speed and Good Quality 12JS 11* 10£ Rapid Speed and Medium Quality Medium Speed and Medium Quality Slow Speed and Medium Quality of. 11* IS* Rapid Speed and Poor Quality Medium Speed and Pool* Quality Slow Speed and Poor Quality Diagram XIX. — Percentage of 1,831 Cleveland pupils found in each of nine groups in speed and quality in silent reading. quality, 12; rapid speed and poor quality, 4; medium speed and good quality, n; medium speed and medium quality, 26; medium speed and poor quality, 12; slow speed and good quality, 4; slow speed and medium quality, 12; slow speed and poor quality, 9. The significant fact brought out by a comparison of these results with those reported in Diagram XIX is that each study emphasizes the same general relation between speed and quality of silent reading. The second method of determining the relation between speed and quality in silent reading was as follows: The average compre- hension score of all pupils reading 100 words in 100 seconds was THE INVESTIGATION OF READING IN A CITY SYSTEM 137 found for each grade. A similar average was found for all pupils in each grade reading 100 words in 90 seconds, 80 seconds, 70 sec- onds, etc. The results are presented in Table XXXIX. The table reads as follows: In the second grade the average compre- hension score for all pupils who required from 98 to 102 seconds to read 100 words was 1 7 ; the average comprehension score for all pupils who required from 88 to 92 seconds to read 100 words was 19. In comparing the average scores made by the selected pupils of the various grades, it should be remembered that second- and third- grade pupils read "Tiny Tad," fourth-, fifth-, and sixth-grade pupils read "The Grasshopper," and seventh- and eighth-grade pupils read "Ancient Ships." TABLE XXXLX Average Quality Scores in Silent Reading of All Pupils Reading at Various Rates Grade IOO Sec. go Sec. 80 Sec. 70 Sec. 60 Sec. So Sec. 40 Sec. 30 Sec. 20 Sec. Sec. I word per sec. 1 . 1 words per sec. 1 -25 words per sec. I.42 words per sec. 1.62 words per sec. 2 words per sec. 2-5 words per sec. 3-3 words per sec. S words per sec. 6-5 words per sec. II Ill IV 17 19 19 21 20 24 23 29 12 13 19 24 28 iS 20 27 10 24 28 15 20 28 17 19 25 3° 17 22 22 19 21 31 34 22 26 28 20 25 30 41 22 24 30 26 29 36 28 14 33 23 28 V VI VII.... VIII . . 18 The table shows that with few exceptions the average quality score for all pupils reading at given rates increases as the rate increases. Some few exceptions to this rule appear in the table. Had more cases been involved, it is believed that most of these exceptions would have been eliminated. It is apparent from the results presented in the table that the rate of reading may become so fast that quality is sacrificed. The average comprehension scores of the most rapid readers in the third, fourth, sixth, and eighth grades illustrate the point in question. This analysis of the relation between speed and quality in silent reading shows very clearly the relative advantage which the more 138 STUDIES OF ELEMENTARY-SCHOOL READING rapid readers possess. If we accept the general principle that rapid rate and good quality usually go together, we are left with the complex problem of explaining why Cleveland pupils who read rapidly on the average are less able to give back what they read than are the pupils in other schools who were tested with the same selections. EXPLANATION OF THE CLEVELAND RECORD The problem with which we are confronted at this point is solved by a consideration of the relative emphasis on speed and quality in the different grades. In Diagram XX rate and quality are both represented by a single curve of progress. The rate of reading is represented along horizontal distances and the quality along vertical distances. Different grades are represented by small circles which appear at various points along the curve of progress. The readjustments in the separate diagrams for speed and for quality which were made necessary by the use of different selections are omitted from these diagrams for the sake of increased clearness and simplicity. Inasmuch as it was impossible to equate a unit of speed as represented in the diagram exactly in terms of a unit of quality, care must be observed in interpreting the curves of progress. As the diagrams are drawn, the most significant points for our consideration are revealed by the changes in the direction of progress which occur from time to time along the curve of progress. The facts in regard to Cleveland and the other cities are pre- sented in Diagram XX. Here we see that the rate of the second grade in Cleveland is nearly 2 words per second, while the rate of the corresponding grade in other cities is about i| words per second. The quality of the second grade in Cleveland, on the other hand, is 25, while that of the same grade in other cities is 30. Following the progress of the Cleveland curve, it is noticed that there is a rapid gain in speed up to the fourth grade, shown by the fact that the curve turns sharply toward the right. The progress is irregular as to quality. There is a relapse to a lower level in passing from the third to the fourth grade. Even after improvement in quality begins in the fifth grade progress is irregular and distinctly slow at first. THE INVESTIGATION OF READING IN A CITY SYSTEM 139 60 1.0 3.0 W: .lson ,o 7 60 \ ,»' 50 3 A, 5 0.. •..6/) / \ ( — 8 iuinc < ' 40 0b£ 2''' erval ion \ Y 1 v', 41 A *••/ ■X y 6 4 / 30 / 3 /1 rs U 20 ■>/ 1.0 2.0 3.0 4.0 5.0 Diagram XX. — Average scores in speed and quality in silent reading in each grade in Cleveland and in thirteen other cities and in three selected Cleveland schools. 140 STUDIES OF ELEMENTARY-SCHOOL READING The curve for the other cities has a wholly different character. During the early years rate improves relatively more than does quality, but there is steady progress in both. In the middle of the elementary course there is a definite change in relation, the upper grades showing greater improvement in quality. This would seem to mean that the mechanical phases of reading must be mastered first, and their mastery is shown by the attainment of satisfactory speed; thereafter the greater emphasis should fall on quality. The whole solution of the problem regarding high rate and deficient quality is now clear. There is not enough attention given to the interpretation of what is read in the grades up to the fourth. There is indeed a high degree of success in perfecting the mechanical operations, as shown in the oral-reading scores and in the rate of silent reading, but the ultimate achievement of the schools is below what it should be in quality because interpretation is not adequately stressed in the lower and intermediate grades. Achievement of various schools. — The scores for the various schools reported in this chapter are based on the records of a limited number of pupils of each grade. While it is felt that the average score for each grade serves as a fairly accurate index of its achievement, it is possible that the average score might have been slightly different had all the pupils been tested. The following comparisons of different schools are not so valuable, therefore, from the standpoint of rigid accuracy as they are suggestive of the kind of analyses which would be productive in every school system. The lower part of Diagram XX shows the results obtained in speed and quality of silent reading in three schools. Fifteen pupils of each grade were tested in the Observation School and in the Wilson School. Seven pupils of each grade were tested in the Quincy School. The diagram shows that the Observation School makes a high record in quality of reading in all grades excepting the fourth. The relatively low score in the fourth grade presents a real problem for solution. The Wilson School makes unusually rapid progress in the second, third, and fourth grades in both speed and quality. Beyond the fifth grade progress is steady and fairly consistent. The peculiar relative achievement of the fourth and fifth grades is unusual and would suggest that the proper amount TEE INVESTIGATION OF READING IN A CITY SYSTEM 141 of emphasis is not given in these grades to problems of silent read- ing. The curve for the Quincy School is very irregular. From the second to the sixth grade inclusive very little progress is made in quality of reading. The comparative study of the records of these schools suggests that there is real need in some schools for a clearer definition of the points to be emphasized in each grade and for a clearer realization on the part of the teacher of the specific means by which the desired goal can be attained. SUMMARY AND CONCLUSIONS The facts brought out in the investigation of reading in Cleve- land were as follows: 1. The average achievement in oral reading reaches a high level for the city as a whole. 2. Variations among schools and among individuals of the same class are very wide. 3. Sex and nationality make for variations in oral-reading achievement. Methods of teaching reading offer very little influ- ence when the city as a whole is considered. 4. In rate of silent reading Cleveland is ahead of thirteen other cities, but behind these cities in ability to interpret what is read. 5. In general, good quality and rapid reading commonly go hand in hand. Poor quality and slow reading are, in like manner, commonly related. The conclusions concerning Cleveland's needs were as follows: 1. More emphasis should be laid on interpretation and relatively less on drill in formal reading. 2. In many schools there is need for a clearer definition of the phases of the work to be emphasized in each grade. 3. In the city as a whole and in each individual school there is need for co-operative effort in careful studies of the problems of teaching reading in order that scientific conclusions might be reached concerning best modes of procedure. CHAPTER VII SPECIAL PROBLEMS IN READING The data secured through the use of the oral-reading and silent- reading tests point toward certain tentative conclusions in regard to many important problems in reading. It is the purpose of this chapter to present the results obtained through a study of each of the following problems: (i) periods of growth in reading achievement; (2) rates in oral reading and in silent reading; (3) relation of growth in oral-reading ability to comprehension of subject-matter read. PERIODS OF GROWTH IN READING ACHIEVEMENT The results of this study are based on data secured by testing 471 pupils and students of the Elementary School, the High School, the College departments, and the Graduate Schools of the Univer- sity of Chicago. The distribution of the persons tested was as follows: 254 elementary-school pupils, 172 high-school students, 27 college students, and 18 graduate students. All the subjects took both the oral- and the silent-reading tests, excepting the first- grade pupils, who took the oral-reading test only. Diagram XXI presents the curve of progress for oral reading. The solid line represents the standard achievement by grades of 2,193 pupils. The broken line represents the achievement of the classes tested for this study. The diagram shows that in general these pupils made scores which are superior to the standard oral- reading scores. The fact that the tests were given early in the year accounts for the apparent relative inferiority of the pupils of the lower grades. Since growth takes place most rapidly in these grades, a difference of two or three months makes a noticeable difference in absolute achievement. The curve of progress may be divided arbitrarily into three parts for the purpose of this discussion. During the first, second, and third grades progress is very rapid. During the fourth, fifth, sixth, and seventh grades progress continues at a regular rate, but by no 142 SPECIAL PROBLEMS IN READING 143 means so rapidly as in the primary grades. At the beginning of the high-school period achievement in oral reading is at a relatively high level. Very little, additional growth takes place during the high-school and college periods. It is in harmony with natural expectation that the first two or three years of school work should be characterized by rapid growth in reading ability. The pupil at this time is devoting a great deal of his attention to the acquisition of reading habits. Every lesson brings the pupil in contact with many new words, and he has I II III IV Col. Grad. Diagram XXI. — Curve of progress for oral reading abundant opportunity to associate the sight of symbols with their proper pronunciation. The words which are learned at this time are usually within the comprehension of the pupil, and word-mastery can progress rapidly, uncomplicated by the problem of mastering meanings. Furthermore, the power of word-analysis develops so rapidly during the second and third grades that the pupil is able to pronounce at sight most of the common words by the beginning of the fourth grade. During the fourth, fifth, and sixth grades silent reading becomes the dominant type of reading. Ability to read is now measured largely in terms of a pupil's ability to master the thought of what 144 STUDIES OF ELEMENTARY-SCHOOL READING he reads. The pronunciation of a very large proportion of the words is more or less familiar. Hence the opportunities for increase in oral-reading ability are much fewer than during the preceding period, and we should naturally expect progress to be less marked. On the other hand, there are factors operating which tend to increase the pupil's ability in oral reading. The subject-matter which is read brings the pupil daily into new fields of thought and in contact with some new words. In order to use these words in class dis- cussion or even in oral reading, the pupil is constantly analyzing words or looking up their pronunciation in the dictionary. Further- more, the pronunciation of partly familiar words becomes more rapid and accurate and the words of a selection are grouped together more economically and effectively. All of these factors result in regular progress in oral reading during the intermediate grades, although the rate of progress is less rapid than in the lower grades. By the time a student reaches the high-school period his reading habits are definitely established. The subject-matter studied at this time repeats for the most part the reading vocabulary which the pupil has already mastered. Technical terms peculiar to certain subjects, proper names, foreign derivatives, etc., are encountered, however, from time to time, and the pupil is forced to continue his study of words. Hence there is some progress from year to year through the high-school and college period, measured largely by the scope of the student's general reading and by the character of his specialized study. Diagram XXII presents the curve of progress for rate of silent reading. The solid line represents the standard rates for 2,654 pupils. The broken line represents the achievement of the classes tested for this study. During the second and third grades the rates at which these pupils read silently approximate the standard rates. From the fourth grade on more rapid progress is made than is indicated by the standard rates. The diagram shows that up to the sixth grade the rate at which these pupils read silently increases very rapidly. Beyond this grade progress is irregular and of no large consequence. The diagram supports in a general way the conclusion reached by Courtis that the curve for careful reading is practically constant from the sixth grade on. This means that SPECIAL PROBLEMS IN READING 145 the rate and character of one's serious reading are fixed in early- school life. During the second and third grades a pupil is developing rapidly in his ability to recognize words at sight. It is in harmony with natural expectation that the rate of silent reading should increase rapidly at this time. As the ability to recognize words increases the power to recognize larger and larger units at a single fixation of the eye develops, with the result that during the intermediate 4.00 3 23 * " 1 1 1 1 y _ __ - • 3.00 P, / / * / • v 1 1 1 li97 / 1 1 ?,,00 1 ?,~\ 497 // 1 1 • 1 1.00 1 1 1 2nd 3rd 4th 5th 6th 7th I II III IV Col. Grad. Diagram XXII. — Curve of progress for rate of silent reading grades the eye moves along the lines at a constantly increasing rate. The maximum development for careful reading is approximated in the sixth grade. Additional points of explanation will be brought out in the following paragraphs. Diagram XXIII presents the curve of progress for quality of silent reading. The solid line represents the standard scores for 2,654 pupils. The broken line represents the achievement of the classes tested for this study. The diagram shows that the scores for these pupils are superior throughout to the standard scores. During the second grade rapid progress is made in quality of silent 146 STUDIES OF ELEMENTARY-SCHOOL READING reading. Progress continues through the third and fourth grades, but it is less marked. From the fifth grade on there is continuous rapid progress in the upper grades and in the high school. It is noted with some surprise that college and graduate students make lower scores than do high-school students. 70 ifljS 1 t 1 ( 1 1 1 37i1 y y s \ \ 60 / 48l2 1 1 1 • 1 i • / 2711 1 y y » *" " 50 ?e;.? i /J T J >: 1711 40 / 28 1 £• // • 1 1 1 1 1 7ll 30 / l/ 18*. 2 1 i 1 1 1 2nd 3rd 4th 5th 6th 7th I II III IV Col. Grad. Diagram XXIII. — Curve of progress for quality of silent reading The rapid improvement in the quality of silent reading at the beginning may be explained in part by the fact that the pupil improves rapidly in his ability to use reading as a tool for getting thought. Habits of attention and study are being developed and refined. The pupil's fund of experience is being rapidly enlarged, with the result that he is able to comprehend the meaning of what he reads more adequately. Some of the irregularities of the quality-curve for the first six years may be explained in part on the basis of the relative emphasis SPECIAL PROBLEMS IN READING 147 which is given to speed and quality. A comparison of the curves for speed and quality shows that during the second grade speed receives relatively less emphasis than is normal, while quality, on the other hand, makes unusual progress. During the third grade speed is emphasized very largely, while there is little progress in quality. During the fourth and fifth years the growth in both speed and quality is normal. During the sixth year quality improves rapidly, but rate increases less rapidly than in the pre- ceding grade. This comparison suggests that improvement in speed and quality of silent reading may be due, in part, at least, to the emphasis which is given to each in the classroom. It is in harmony with natural expectation that the quality of one's reading should improve very distinctly during the fourth, fifth, sixth, and seventh grades. It is just at this time that the pupil learns to use reading in a large way as a tool to enlarge his world of meanings. He reads copiously and with a purpose. The natural result should be increased power in getting the thought of the printed page. As the pupil continues through the grammar grades and in the high school, his methods of study are improved. College and graduate students show less power than high-school pupils, according to the diagram. The explanation for this appar- ent decrease in reading efficiency lies largely in the fact that adults are much more self-conscious during a test than are high-school pupils. The following interesting facts were brought out as a result of observations and questioning which were carried on by the writer: first, adults are more conscious of the significance of tests than are high-school pupils; second, they are less accustomed to study under the immediate observation of others, and hence feel more confused; third, tests are found to be a severe trial in a very large proportion of the cases. These facts justify the acceptance of the above-mentioned explanation for the apparent decrease in reading efficiency of college and graduate students. COMPARISON OF RATES IN ORAL READING AND IN SILENT READING The subjects for this study included the 471 pupils and students of the University of Chicago mentioned in the preceding study. The subject-matter upon which the rate of oral reading is based is 148 STUDIES OF ELEMENTARY-SCHOOL READING paragraph 3 of the oral-reading scale. The selections upon which the rate of silent reading is based are the selections of the standard silent-reading tests. For the purpose of checking the relative diffi- culty of "Tiny Tad" and paragraph 3 a class of third-grade pupils were asked to read them orally. The results showed that they were practically equal in difficulty for oral reading. The average time required by the various grades to read paragraph 3 was determined and the rate was expressed in terms of the number of words read per second. The results by grades were as follows: First Second Third Fourth Year Year Year Year High High High High IA IIA IIIA IVA VA VIA VIIB School School School School Col. Grad. 1. 00 1.86 2.45 315 351 3-8o 3.90 3.90 4.10 4.00 3.98 4.02 3.80 The solid line in Diagram XXIV represents the curve of progress for rate of oral reading. The significant point about the curve is y 4.00 • ^ s y / "^ ^\ / / y> 1 3.00 1/ /l f 2.00 // // ' 1.00 1st 2nd 3rd 4th 5th 6th 7th I II III IV Col. Grad. Diagram XXIV. — Comparison of oral- and silent-reading rates of 471 pupils and students in the University of Chicago. that progress continues throughout the grades and reaches a maxi- mum rate during the high-school period. The increase in rate during the lower grades is due to the fact that words are recognized at sight with increasing rapidity and groups of words are combined into more effective units for oral expression. The very high rate SPECIAL PROBLEMS IN READING 149 during the high-school period may be accounted for in part by the rapid offhand way in which high-school students do most easy tasks. The decrease in the rate of graduate students is due to the greater deliberateness which characterizes older people. The dotted line in Diagram XXIV represents the curve of progress for rate of silent reading. The diagram shows that the rate of silent reading for the second and third grades is less rapid than the rate of oral reading. This is significant in view of the fact that the two selections were read orally at the same rate by a group of pupils of corresponding ability. It indicates that when a pupil is asked to read a selection for the content he may read it more slowly and more carefully than if he were reading it orally without directing special attention to the content. During the fourth, fifth, and sixth grades pupils read silently at a rate which is dis- tinctly higher than the oral-reading rate. The fact that the rate of careful silent reading exceeds the rate of oral reading needs explanation. The ordinary child who has reached the fourth grade has had a great deal of training in oral reading, with the result that he can now use reading independently. As has been pointed out in previous paragraphs, he begins to read more rapidly than in the primary grades. He becomes interested in the subject- matter and his eyes run along the lines at a rate which surpasses the rate of vocalization. As a result we find that the rate of silent reading exceeds the ordinary rate of oral reading in the intermediate grades. It is interesting to note, on the other hand, that after the fourth grade the rate of silent reading parallels very closely the rate of oral reading. This fact suggests that the habits of oral reading which receive so much attention in the schoolroom today may have a retarding influence upon the rate of silent reading. RELATION OF ORAL-READING ACHIEVEMENT TO COMPREHENSION OF SUBJECT-MATTER READ The question has frequently been raised : Is it possible to develop so rapidly in ability to pronounce words that one is able to read orally subject-matter which is beyond his comprehension? This study was undertaken to determine a partial answer at least to this question. I50 STUDIES OF ELEMENTARY-SCHOOL READING During the winter of 1914 the oral-reading test was given to the pupils of the Elementary School of the University of Chicago. Four questions were carefully prepared for each paragraph. The questions for paragraphs 3, 7, and 11 were as follows: Paragraph 3 1. Where did the king and queen live ? 2. Why were they unhappy? 3. What did they find at their door one day ? 4. What did they do with the little boy and girl ? Paragraph 7 1. What time of day was described in the paragraph ? 2. Behind what had the sun gone down? 3. Describe the sky as it looked on this evening. 4. What is meant by "The dull mountains stood in contrast against the western sky"? Paragraph ii 1. What is meant by "The attractions of the American prairies have been overcome by the azure skies of Italy" ? 2. What are antiquities of Roman architecture ? 3. What are architectural studies? 4. What is meant by "My delight in these studies verges on to a fanati- cism"? The pupils read the successive paragraphs of the oral-reading test according to the usual directions. After the reading of each paragraph the pupil was asked the questions which covered that paragraph. A grade of 25 per cent was given for each question answered correctly and a grade of 12I per cent for a partial answer. A record was made of the average score of the pupils of each class for the first selection of the test and of the average score of the class for the hardest selection which each pupil was able to read success- fully. Four weeks later the test was given again. At this time the test was preceded by a fifteen-minute study-period. The following directions for the study-period were given: "You are going to take the oral-reading test again. Study the paragraphs carefully for 15 minutes so that you can read them as well as possible and so that you can answer any questions which might be asked about each paragraph." SPECIAL PROBLEMS IN READING I5i The solid line in Diagram XXV represents the average achieve- ment of each of the grades from IIB to VIIB inclusive for the first reading, and the dotted line represents the average achievement by grades for the second reading. The point revealed by the diagram, which is significant for the present discussion, is that the improve- ment in oral-reading achievement which resulted from fifteen minutes of study and from the natural growth in reading ability Diagram XXV. — Influence of a fifteen-minute study-period on achievement in oral reading. during one month was practically uniform throughout the grades. As a result of this uniformity any striking differences in comprehen- sion scores will measure with a fair degree of accuracy the relation of the increase in oral-reading ability to the comprehension of what is read. Diagram XXVI shows the influence which the increase in oral- reading ability had upon the comprehension of what was read. The curve which represents the scores made on the easiest paragraph 152 STUDIES OF ELEMENTARY-SCHOOL READING without study shows that the pupils of the second grade were able to answer the questions on the first paragraph practically as well as were pupils in the upper grades. Some improvement resulted from a careful study of the paragraphs. This was uniform, but not striking. Inasmuch as the paragraph was easy and the thought 100$ 90 80 70 €0 50 40 30 20 J, < s. »«. * S- y - - \ / y y A; I \" / \ y'. .— " ;: / \ V \ % \ V * * * " ^ * \ 1 "—/ ^y t 1 1 t / 2B 2A 3B 3A 4B 4A 5B 5A 6B 6A 7B Easiest paragraph without study — Easiest paragraph with study Hardest paragraph without study -- Hardest paragraph with study Diagram XXVI. — Comprehension of the subject-matter read as related to the mastery of mechanical difficulties. simple, the pupils of the lower grades were able to understand its meaning practically as well as the pupils of the upper grades. The curve which represents the comprehension scores for the hardest paragraphs read without study by pupils in the second grade drops very little below the curve for the comprehension scores for SPECIAL PROBLEMS IN READING 153 the easiest paragraph. These high scores in comprehension are easily explained by the fact that, although the pupil makes rapid progress in oral reading at this stage, he does not encounter many words whose meanings are not already familiar to him because of their frequent use in everyday conversation. Hence if the word is recognized successfully, its meaning is understood immediately. During the third and fourth grades the comprehension scores drop surprisingly. The continued rapid progress in oral reading during these grades causes the pupil to encounter a large number of new words whose meanings he has not mastered. His power of word- analysis enables him to pronounce the words successfully, but his ability to interpret the meaning of what he reads lags behind. During the fifth, sixth, and seventh grades the pupil reads widely. His world of meanings is greatly enlarged. He has become more highly trained in discovering meanings from the context as he reads. Hence we note that the curve of comprehension rises gradually during these grades in spite of the fact that oral-reading ability has been increasing constantly. The curve for the comprehension of the hardest paragraph read with study shows some very interesting facts. Pupils of Grade IIB make a better score on the hardest paragraph read with study than on the easiest paragraph read without study. This is explained by the fact that their power to pronounce words independently at this time is so limited that it does not bring them in contact with many words whose meanings are unfamiliar. Hence the study of the paragraphs enabled the pupils to master the meaning of the hardest paragraphs more fully than they had mastered the meaning of the easiest paragraph through sight reading. Pupils of Grades HA, III B, and III A have developed considerable ability in independent word-analysis. As a result their study of the paragraphs enabled them to read successfully paragraphs which were much beyond their comprehension. The effect of this is shown by the rapid decline of the comprehension-curve. By questioning each pupil carefully concerning his method of study during the study-period it was found that most of the pupils spent the greater part of their time analyzing words in order to pronounce them. The pupils in the grades beyond the third, however, reported more and more 154 STUDIES OF ELEMENTARY -SCHOOL READING emphasis given in their study to the meaning of what they read. This change in the direction of their attention, together with their increasing background of facts and meanings, enabled them to secure relatively higher scores in comprehension than they had made on easier paragraphs without study. The foregoing discussion leads to the conclusion that it is possible to advance so rapidly in ability to pronounce words that one's ability to comprehend the more difficult passages which he can read successfully is relatively decreased. This is particularly true if attention is directed to pro- nunciation of words rather than to meanings. CONCLUSIONS The tentative conclusions which have been reached as a result of the various investigations reported in this chapter relate only to the school in which the investigations took place. Whether or not these conclusions relate to all schools is a problem for further investigation. The universal value which these studies have is that they point out the presence of significant problems connected with the teaching of reading. As a result of the objective study of read- ing in more than two hundred schools, it is believed that there are abundant opportunities for productive studies of reading along similar lines in each classroom. INDEX Abeil, A. M., 14, 24, 26. Achievement : expressed in numerical terms, 43; growth periods in, 142-47. Administration of reading tests, 28. Adopted values, 93-94. Aldine method, 127. American schools, 129. Ancient Ships, 50. Anderson, H. W., 24, 29. Answers to questions, 105, in. Assimilation, 22. Average quality marks, 54. Average rates, 54. Average scores in oral reading, 121, 124, 125. Ayres, L. P., 9, 82, 118. Beer, Max, 22, 26. Boggs, L. P., 26. Bowden, J. H., 22, 23, 26. Breathing, relation of, to oral reading, 27. Brown, H. A., 15, 17, 18, 19, 26. Brownell School, 124. Buckingham, B. R., 71, 83. Careful reading, rate of, 20. Case-Woodland School, 124. Check card (phonetic), 3. Class scores, 40, 42. Cleveland schools, 75, 80, 119. Cleveland survey, 14, 28, 118. Comparison of comprehension in oral and silent reading, 25. Comparison of rates: arithmetic mean and harmonic mean, 16; oral and silent, 25, 147. Comprehension: basis of, 105; direc- tions to pupils, 14; during- oral and silent reading, 25; growth in, 25; methods of scoring, 16; methods of testing, 16. Correlation: between oral-reading achievement and comprehension of subject-matter read, 149; between speed and comprehension of silent reading, 24-25, 131-38; between two records for rate, 105. Courtis, S. A., 11, 12, 13, 14, 15, 19, 20, 21, 26, 27. Curve of progress: Oral reading, 45, 143; quality of silent reading, 54, 146; rate of silent reading, 53, 145. Dearborn, W. F., 21, 22, 27. Distribution of reading ability, 82. Eagle School, 124. Educational investigations, Yearbook of, 27. English, standard tests in, 26. Equality of differences, 63. Errors in oral reading, 35. Eye co-ordination, 80. Eye movements, habits of, 21. Failures by grades, 81, 82. Familiarity, influence of, 21. Freeman, F. N., 17. Gill, E. J., 27. Gilliland, A. R., n, 18, 27. Grading reading errors, 78. Grand Rapids Survey, 47. Grasshoppers, The, 49. Gray, C. T., 27. Gross errors, 78, 79. Growth periods in reading achievement, 142-47. Hebrew schools, 129. Hendricks, E. L., 21, 24, 25, 28. Huey, E. B., 21, 22, 28. Illinois schools, 74, 80. Individual records, 38, 39. Individual score sheet, 41. Individual scores, 40. Individual variations, 123. Influence on oral-reading progress: of fifteen minutes of study, 151; of method, 12S; of nationality, 129; of sex, 127. Initial oral-reading study: method of selecting paragraphs, 61; purpose, 60; reorganization of paragraphs, 65; scal- ing paragraphs, 69; school population tested, 68; scoring results, 68; steps of difference between paragraphs, 64. Inner speech, 29. Insertions, 35, 79. Investigations of reading, 28. Italian schools, 129. Jones, R. G., 2, 28. Judd, C. H, 28, 108, 118. 155 i56 STUDIES OF ELEMENTARY-SCHOOL READING Kansas Silent-Reading Test, 8, 29. Kelly, F. J., 8, 9, 16, 29. Kentucky School, 124. Learning to read, 26. Lip movements, 21. McLaughlin, K., 105. Materials for reading tests, 28. Mead, C. D., 14, 24, 25, 29. Measurement of ability to read, 26. Measurement of efficiency of instruction, 26. Measuring-scale for oral reading, 92; validity of, 93. Mechanics of elementary reading, 2, 28. Messmer, 0., 22, 27. Methods of teaching reading, 27. Minor errors, 78. Monosyllabic words, 22. Normal distribution of reading ability, 83. Normal frequency curve, 89. Normal probability integral, 85. Normal reading, rate of, 20. "Normal surface of frequency," 70. Numerical score, 44. Oberholtzer, E. E., 12, 13, 14, 16, 19, 22, 23, 29.. Observation School, 139. Omissions, 35, 79. Oral reading: comprehension of, 25; improvement of, from grade to grade, 90; loss in rate of, during summer, 24, 29; measure for, 42; measurement of , 27; ranks compared, 37; rate of, 23; relation of breathing to, 27; tests of, 28. Oral reproductions, 107. Outhwaite School, 125. Packer, P. C, 24, 29. Paragraphs: methods of selecting, 61; second series of, 73; third series of, 74; P. E. equivalents, 69. Pintner, R., 15, 16, 23, 24, 25, 29. Practice in reading, influence of, 22. Practiced readers, 30. Probable error, 70, 83. Probability integral, 85. Purpose of thesis, 1. Quality of silent reading: method of determining, 54; relation to speed, 24; standard scores, 54; validity of stand- ards, 112. Quantz, J. A., 21, 22, 24, 30. Quartile deviation, 70. Quincy School, 125, 139. Rankings of paragraphs, 62. Rapid readers: achievement of, 134, I 36, 137; characteristics of, 20; superiority of, 24. Rapid reading, advantages and methods, 26. Rate of silent reading, 19, 23; careful, 20; directions to pupils, 14; factors influen- cing, 21-23; influence of practice on, 20, 22; methods of computing, 15; normal, 20; ordinary, 28; progress in, 20; relation to quality, 24; standard rates by grades, 19; standards in, 19, 27; terms in which expressed, 15; validity of standards, 112; very rapid, 28._ Reading achievement, index of, 18. Reading errors: description of, 35; grading of, 78. Reading investigations: importance of, 118; summary of, 10. Reading, oral and silent distinguished, 28. Reading paragraphs: initial series, 65- 68; second series, 73; standardized, 46; third series, 74. Reading tests, 28; administration of, 28; materials for, 28. Reading unit, 18. Reading vocabulary, 30. Real values in oral-reading scale, 94. Relation between rate and quality, 131. Relative ranks of schools, 94. Reliability of a single record for rate, 105. Repetitions, 35, 79. Reproductions: group-idea method, 108; methods of scoring, 17, 108; oral, 107; word-counting method, no; written, 107. Revised score sheet, 55. Rice School, 125. Romanes, G. J., 30. Rubins, R. B., 30. Ruediger, W. E., 22, 30. Scaling paragraphs, 61, 69, 70; method of, 81. Schmitt, C, 30. School population tested, 74, 100, 112, 119. Scoring oral-reading tests, 68, 76. Sholty, M., 30. Silent reading: quality of, 54; rate of, 19, 23 ; relation between speed and quality, 28; speed of, 53; tests of, 28. Silent-reading tests, the, 47; criticism of adopted methods, 103; directions for giving, 51, 52, 58; directions for scor- ing, 59; form in which printed, 103; length of, 104; limitations of, 115-17; preliminary study, 99-103; results of, INDEX 157 130; subject-matter of, 48; validity of, 99- Slow readers: achievement of, 134, 136, 137; characteristics of, 20. Standard oral-reading test: conditions under which given, 32; directions for giving, 34; graphical representation of scores, 44, 45; limitations of, 96-98; scope of, 35; standards for scoring, 39; tabulation of data, 38; test paragraphs, 33; validity of, 60. Standard scores: oral reading, 45, 47; quality of silent reading, 54; rate of silent reading, 54. Standard tests in English, 26. Standard vocabulary, 28. Standardized reading paragraphs, 46. Standards in rate of reading, 27. Standards of teaching efficiency, 31. Starch, D., n, 13, 15, 18, 19, 25, 30, 104. "Steps of difference" between para- graphs, 64. Steps of difference in difficulty, aver- age, 89. Substitutions, 35, 79. Surface of distribution, 123. Surface of frequency, 70. "Surface of normal distribution," 83. Tabulation sheet, 38, 39. Testing reading, methods of, 28. Tests: by whom given, 14; criticism of standard tests, 2-10; directions con- cerning speed and comprehension, 14; duration of, 12; form of presentation, 12; indicating portions read, 13; Kansas silent-reading, 8; number of selections used, 11; standards in mechanics of elementary reading, 2; subject-matter for, 11; Thorndike's visual vocabulary scale, 4; under- standing of sentences, 6. Tests for expression, 37. Thorndike, E. L., 2, 4, 6, 8, 9, 27, 31, 70. Tiny Tad, 48. Training testers, 120. Turner, E. A., 9, 31. Types of errors, 60, 126. University Elementary School, 142. Valentine, C. W., 31. Validity of standards, 68. Variations in schools, 124. Visual perception, 22. Visual vocabulary, scale for, 4. Vocabulary test, 28. Waldo, K. D., 11, 12, 13, 14, 16, 18, 20, 25,3i- Ward method, 128. Weighting of errors, 78-80. Wilson School, 139. Written reproductions, 107. Yule, G., 16. BD -S«| 4?\ 0' •V . ^ > ^ vP

^ ***** :% M^'' ^ "'ife 4 °^