THIS IS AN ORIGINAL MANUSCRIPT IT MAY NOT BE COPIED WITHOUT THE AUTHOR’S PERMISSION A LABORATORY STUDY OF THE READING HABITS OF SPANISH-SPEAKING CHILDREN Approved: THIS IS AIT ORIGINAL MANUSCRIPT IT MAY KOT BE COPIED WITHOUT. TELE AUTHOR’S PERMISSION Approved: Dean of the Graduate School. A LABORATORY STUDY OF THE READING HABITS OF SPANISH-SPEAKING CHILDREN THESIS Presented to the Faculty of the Graduate School of The University of Texas in Partial Fulfillment of the Requirements For the Degree of DOCTOR OF PHILOSOPHY By James Knight, 8.5., M.A. Austin, Texas June, 1931 330868 PREFACE The eye-movement technique has been used in several studies of the reading of foreign languages. The present research, however, is believed to be the first application of the technique to a study of the reading habits of children who learn to speak one language and have their first, and in many cases all of their subsequent reading instruction in another language. Spanish-speaking children were chosen as subjects for the research for two reasons: first, because children of this language group who fulfill the conditions of the study are found in large numbers in the city of Austin; and second, the choice of these children is in line with other researches dealing with the education of Spanish-speaking children which are now being carried on in the University of Texas. It is a pleasure to acknowledge my indebtedness to the many parsons who have so cheerfully assisted me in carrying out the study. I am indebted to Professor H. T. Manuel of the Department of Educational Psychology for directing the preliminary testing program for the selection of subjects, and for invaluable advice and assistance throughout the study. I am under lasting obligations to the following named principals of Austin schools and to their teachers and pupils for the cooperation they have given me in securing my data: Miss Josephine Houston, Fulmore School; Sister Leonidas, Our Lady of Guadalupe School; J. B. Watkins, John T. Allan High School; and W. L. Darnell, Palm School. Mr. B. F. Holland of the Department of Educational Psychology was kind enough to operate the eye-movement camera for me, which left me free to give all of my attention to the subjects of the study while the eye-movement records were being made. I am indebted to my Supervisory Committee for reading the manuscript and for offering helpful suggestions leading to its improvement. Finally, I am obligated to Professor 0. T. Gray, the director of the research, who has given material aid, helpful suggestions, and unfaltering encouragement from the production of the first eye-movement record to the writing of the final conclusion. I wish to express my sincere thanks to all who have assisted me. They are entitled to a share of the credit for whatever is valuable or worthwhile in the study, the author alone is responsible for its shortcomings. James Knight Austin, Texas June, 1931 CONTENTS Chapter Page I Statement of the Problem and Method of Approach . 1 II The Use of the Eye-Movement Technique in the Investigation of Reading Habits 27 III Experimental Findings and Results . . 44 IV Individual Analyses and Comparisons . . 78 V Relationship Found Between Various Measures of the Reading Habits of Spanish-Speaking Subjects .... 141 VI Interpretations and Conclusions . . . 154 Bibliography 168 Appendix 169 LIST OF TABLES Number Page I Mean Reading Ages and Picture Supple- ment Scores of the Groups Used ... 17 II Mean Arithmetical Computation Ages, Mental Indices, and Mental Ages ... 18 111 Unstandardized Measures of Rate, Comprehension, and ’’Carelessness” (Group Means) 20 IV Summary of the Records of a Mature and an Immature Reader 34 V Number of Subjects Comprehending Each Selection 46 VI Average Time Per Line ....... 48 VII The Reliability of the Differences Between the Groups for Average Time Per Line on Four Selections . . 50 VIII Average Time Per Line for Eight- Year-Old Subjects ....... 53 IX Average Number of Pauses Per Line ... 54 X Average Number of Fixations Per Line (Eight-year-old Subjects) 56 XI Average Duration of Fixations .... 57 XII Reliabilities of Group Differences in Average Duration of Fixations . 58 XIII Average Number of Regressive Move- ments Per Line ........ 62 XIV Regressive Movements of the Eight- Year-Old Groups 63 XV Number of Regressive Movements Resulting from Failure of Return Eye-sweep from End of One Line to Carry Back to Beginning of the Line . . 65 Number VII XVI Length of Pre-Regressive Fixations Compared with Regressive Fixations . . 67 XVII Variability in Eye-Movement Performance 69 XVIII Comparison of Reading Performance in Spanish and English ...... 71 XIX Ranks of Groups with Reference to Va,rious Measures 75 XX Summary of Records of Subjects TGSB and WBEB 88 XXI Test Records of Subjects Used in Making Individual Comparisons .... 90 XXII Summaries of Eye-Movement Records for Three Pairs of Spanish-Speaking Subjects, Ages 11 and 14 ... . 100 XXIII Comparison of the Records of an Eleven-Year-Old Spanish-speaking Subject With the Records of English-speaking Subjects Who Have the Same Reading Age 103 XXIV Comparison of the Different Groups of Subjects With Reference to Special Word Difficulties 112 XXV Summary of Group Performances on the Last Line of Selection I 116 XXVI Difference In Eye-Movement Records for Spanish-Speaking Subjects for Supplying a Word Which Has Been Left Out at the End of a Sentence . . 127 Number Page XXVII Rank Difference Intercorrelations Between Eye-Movement Measures for Three Selections of Increasing Difficulty ........ 147 XXVIII Relationship Between Test Measures and Eye-Movement Measures . . . 151 LIST OF PLATES Number Page I Silent Reading by Subject CASI 4 of Selection 111 31 II Silent Reading by Subject SSA of Selection 111 32 111 Silent Reading by Subject CSSII of Selection 111 . 37 IV Silent R e ading by Subject BGSI4 of Selection IV 41 V Silent Reading by Subject OSSII of Selection IV 42 VI Silent Reading by Subject ECSB ... 80 VII Silent Reading by Subject MFEB ... . 83 VIII Silent Reading by Subject TGSB . 86 IX Silent Reading by Subject WBEB . 87 X Silent Reading by Subject JRSI4 of Selection IV 91 XI Silent Reading by Subject CLSII of Selection IV . . . 92 XII Silent Reading by Subject PCSI4 of Selection II 95 XIII Silent Reading by Subject OPSII of Selection II 96 XIV Silent Reading by Subject SHSI4 of Selection 111 98 XV Silent Reading "by Subject TCSII of Selection 111 99 Number X XVI Silent Reading by Subject ACSII of Selection 111 105 XVII Silent Reading by Subject LJEII of Selection 111 106 XVIII Silent Reading by Subject ESEB.3 of Selection 111 107 XIX Silent Reading by Subject CSEI4 of Selection 111 108 XX Silent Reading by Subject CANES.B of Selection 111 109 XXI Silent Reading by Subject JDEI4 of Selection 111 110 XXII Example of a Ready Response at the End of a Selection 115 XXIII Example of Difficulty at the End of a Selection 115 XXIV Fairly Mature Performance at the End of a Selection . 115 XXV Less Mature Performance at the End of a Selection . 115 LIST OF FIGURES Number XI 1 Number of Fixations for Last Line of Selection I — Spanish 14 and English 14 118 2 Number of Fixations for Last Line of Selection I — Spanish 11 and English 11 118 3 Number of Fixations for Last Line of Selection I — Spanish 11 and 14 and English 9.7 121 4 Number of Regressive Movements for Last Line of Selection I — Spanish 14 and English 14 ...... 123 5 Number of Regressive Movements for Last Line of Selection I — Spanish 11 and English 11 123 6 Number of Regressive Movements for Last Line of Selection I — Spanish 11 and 14 and English 9.7 . . . . 123 7 Average Duration of Fixations — Last Line of Selection I — Spanish 14 and English 14 .......... 125 8 Average Duration of Fixations — Last Line of Selection I — Spanish 11 and English 11 125 9 Average Duration of Fixations — Last Line of Selection I — Spanish 11 and 14 and English 9.7 125 10 Average Number of Fixations Per Line — Spanish 14 and English 14 . 131 11 Average Number of Fixations per Line — Spanish 11 and English 11 131 Number XII 12 Average Number of Fixations per Line — Spanish 11 and 14 Compared with English 9.7 132 13 Average Duration of Fixations — Spanish 14 and English 14 134 14 Average Duration of Fixations — Spanish 11 and English 11 134 15 Average Duration of Fixations — Spanish 11 and 14 and English 9.7 135 16 Average Number of Regressive Movements per Line — Spanish 14 and English 14 ... 137 17 Average Number of Regressive Movements per Line — Spanish 11 and English 11 ... 137 18 Average Number of Regressive Movements per Line — Spanish 11 and 14 and English 9.7 137 19 Relative Standing with Respect to Average Time per Line for Three Selections of Increasing Difficulty — Twelve Spanish- Speaking Subjects 144 20 Average Number of Fixations per Line of Three Selections of Increasing Difficulty — Twelve Spanish-Speaking Subjects . 145 21 Average Duration of Fixations for Three Selections of Increasing Difficulty — Twelve Spanish-Speaking Subjects 145 33 Average Number of Regressive Movements per Line for Three Selections of Increasing Difficulty — Twelve Spanish-Speaking Subjects 146 CHAPTER I STATEMENT OF THE PROBLEM AND METHOD OF APPROACH A LABORATORY STUDY OF THE READING HABITS OF SPANISH-SPEAKING CHILDREN Introduction The 1920 census showed that something over thirteen million people in the United States were foreign born. Many of these people speak a language other than English in their homes, and in some localities the foreign language may not be lost for two or even three generations. Children from such homes, in many instances, enter school with a speaking knowledge of the home language but little or no knowledge of English. Very often they use English only at school and only in rare cases do they have any contacts with written materials in their native language. It is quite generally admitted that such children have a language handicap. Obviously, if children have to learn the language of the school as well as the traditional material, they have more to learn than English-speaking children. In many cases they will be unable to understand just what kind of responses they are expected to make to the school situation. Because they have such elementary language adjustments to make, the content of their school work will often be below their mental capacities. Some investigators have gone so far as to say that these bilingual children have a dual language handicap* That is to say, they do not show normal growth in either language. tested two small groups of bilingual Jap- anese children in California on parallel forms of the National Intelligence Test, one form in English and the other in Japanese. Both groups fell below the norms for the United States and for Japan, but the older ones were more successful with two languages than the younger ones. The same author summarizes the literature on this point from European studies. Saer* 5 and also Smith found that Welsh children were handicapped under Welsh and English instruction. Rouget, 4 studying French children K under French and German instruction, and Pavlovitch, studying Servian children under Servian and French instruc tion, found no language handicap. Purpose of the Present Research It is the purpose of the present research to investi- gate some phases of the development of the ”second language.” The particular problem selected relates to reading, and the eye-movement technique was chosen as a means for getting the measures which are essential to such a study. Specifically stated, the purpose of the present research is to make an analysis by means of the eye-movement technique of the reading habits of children who learn to speak one language at home and who are taught to speak and read another language at school by the direct method of instruction. Choice of an Experimental Group This study could be carried out in any country or in any school where children are found whose home language is different from that of school instruction. Favorable situations for such a study exist in southwestern United States where there are literally thousands of children who fulfill the conditions mentioned above. These children come from homes where Spanish is the spoken language and they have all of their school instruction in English. As a rule their knowledge of English is very limited at the time they enter school. Consequently, these children furnish a very convenient experimental group. Actually then, although the findings of the study should be to some degree applicable to any language group, the present investigation is a laboratory study of the reading habits of Spanish-speaking children. Studies of the Spanish-Speaking Child The Spanish-speaking child has been the subject of many educational studies during the last decade. In most cases the investigations have been based upon the results obtained through the use of standardized tests, and in many instances the showing made by the Span ish-speaking child depended upon his ability to read English. Some studies have been centered entirely upon his reading ability, but as yet, so far as the writer has been able to learn, no laboratory study of his fundamental reading habits has been made. In fact, no record has been found of any laboratory study of the reading habits of children who come from homes where a foreign language is spoken to the typical American public school where only English is used. The Language Handicap of the Spanish- Speaking Child Turning to studies of the language handicap with special reference to the Spanish-speaking child, we find 6 Garretson concluding at the end of an investigation of the causes of retardation among Mexican children that the language difficulty is an important factor in grades one and two, but of decreasing importance in the intermediate and upper grades. Throughout the elementary grades, however, the Spanish-speaking child does better in subjects where little reading is required. Manuel*s conclusions based on materi al gathered from ten school systems bears this out: ’’When tested by commonly used measures of intelligence and achievement Mexican children on the average fall below comparable age and grade groups of other white children. The standing in arithmetical computation and in spelling is apt to be somewhat more favorable, sometimes equaling or exceeding that of other white children in the same grade.” 7 The Progress of the Spanish-Speaking Child in Reading used the Monroe Standardized Reading Test, Revised, and tested 6,119 English-speaking children and 3,356 Spanish-speaking children in the schools of New Mexico for comprehension and rate in silent reading. These children were in grades three to eight. The Spanish-speaking children tested below the English-speaking children in both speed and comprehension and a distinct tendency was shown for the differences between the two language groups to grow progressively larger from grade to grade. The same test was used in a study of reading in the Harlingen, Texas, Public Schools. $ Here 327 English- speaking and 82 Spanish—speaking children .were tested in grades three, four,and five. The English-speaking group made a better record on comprehension, but their scores did not get progressively higher. The Spanishspeaking children made the more favorable showing with reference to rate in the third and fifth grades. Very similar results are reported from the use of the Thorndike-McCall Reading Test with 90 English-speaking and 223 Spanish-speaking children in the Brownsville schools*For low and high sections of the fourth and fifth grades the English-speaking groups make higher scores except in the low fifth grade. The Use of the Laboratory Method in the Investigation of Reading Habits in a Foreign Language While no laboratory investigation has been made of the reading habits of children who learn to speak one language at home and have all their school work in another language, a closely related study has been made. Buswell,using the. eye-movement technique, made laboratory analysis of the reading of modern foreign languages by elementary school pupils, high school and college students. He found that high school and college students taught by the direct method approach adult reading habits by the end of the second year of study. Elementary school pupils are as a rule unable to do this. The subjects chosen by Buswell were students of a foreign language who took up the second language after they had learned to speak, read, and write English with ease and fluency. During the period in which they studied the foreign language, English was the language of their other classwork and of their home and recreational contacts. Buswell’s results throw little or no light on the present problem, but his technique is suggestive. The question arises as to what the application of the same technique will reveal in the case of children who learn to speak one language and later attend school where all their oral and written instruction is given in another language. Description of the Technique The eye-movement technique has been used in numerous studies at the University of Texas and at other universities, as the University of Chicago. It has been described so many times that little needs to be said of it here in a descriptive way. Briefly, the subject reads before a specially designed camera which carries standard motion picture film. The material to be read is so placed that light reflected from the subjects eye is thrown into the camera in such a manner as to make a photographic record of all his horizontal eye-movements. The light which is thrown upon the subject’s eye is cut twenty-five times a second by an electrically driven tuning fork. As a result, a pause in reading is shown on the film as a series of dots running lengthwise of the film and parallel to its edges. Forward sweeps of the eye are usually too rapid to register, but the location of the next pause is indicated by another series of dots to the right of the previous line of dots. Backward or regressive movements are indicated by a line of dots to the left of the preceding line. The developed films are later run through a projection lantern and thrown on a screen which has a copy of the material read fixed to it. Thus it is possible to tell the position of each fixation or pause made by the subject, the number of pauses he makes per line, the number of regressive movements he makes, and, since the period of vibration of the tuning fork is known, to tell the duration of each fixation in twenty-fifths of a second. Specific Questions to Which Answers Are Sought Some of the questions which may be raised and which the present study will attempt to answer are: 1. When compared with English-speaking children of their own age and reading ability as measured by standardized tests, do Spanish-speaking children show more or fewer fixations or pauses per line in reading any given selection with comprehension? 2. Do they spend more or less time at each fixation? 3. On the average, is their total time greater or less for reading a line or a selection? 4. Do they make more or fewer regressive movements? 5. Which group shows the greater variability in performance? B. Which group shows the greater change in performance with a change in the difficulty of material read? The same questions and comparisons will be applied with reference to another English-speaking group with the same reading age, but two to five years younger in chronological age. Finally, some light may be thrown on such general reading problems of the Spanish-speaking groups as the following: 1. How do they improve in reading habits from one age level to another? 2. At what age do they approximate adult perform' anoe? 3. What effect does rhetorical form have upon reading performance? For example, what effect does the occurrence of figures of speech in the material read have upon the performance of a subject? 4. Can any distinctive reading habits be pointed out which are unique to or characteristic of the Spanishspeaking child? 5. What objective measures furnish the best basis for predicting the eye-movement habits of Spanish-speaking children? How the Present Investigation Differs From Previous Studies There are several ways in which the present investigation differs from previous studies of Spanish-speaking children. One difference to be noted is the fact that while in previous studies their performances were usually compared with those of children of the same grade placement, in the present study their performances are compared with those of children of the same age and of the same ability in reading as measured by commonly used tests, and with another group where ability but not age is held constant. In the present study, also, the eye—movement technique is made the basis of the investigation after the testing technique has been used to pair the various groups with reference to ability. Subjects Used The subjects chosen for study were Spanish-speaking boys who had little or no knowledge of English when they entered school and who have learned English by the direct method. These subjects were selected on three age levels, eight, eleven, and fourteen, on the basis of performance on standardized reading tests. Originally it was planned to have six subjects for each level, but one record of a subject on the eight-year level was not usable due to excessive head movement while reading. Each Spanishspeaking group was paired with an English-spesk ing group of the same age and reading ability as measured by standardized tests. Since the Spanish-speaking children are, as a rule, about two years retarded in their school work, they were necessarily paired with English-speaking children of the same degree of retardation when age is held constant. This condition is particularly operative on the eleven and fourteen year levels. The question is immediately raised as to whether typical results can be obtained from a comparison of retarded Spanish-speaking children with retarded English-speaking children. The former group may be normal with respect to intelligence, and retarded because of a language handicap, while the English-speaking group may be retarded primarily as a result of inferior learning ability. So, in order to offset any misleading results that might arise from such pairings a second conirol group of twelve English-speaking children was selected whose mean chronological age was 9.7 years and whose mean age was approximately the same as that of the eleven-and fourteen-year-old Spanish-speaking groups taken together. Photographic records were taken of three adult readers. One of these adults was a graduate student of Spanish-American extraction who has about equal facility in reading Spanish and English. One of the other adults was a graduate student and the other a member of the University faculty. The last two records were taken to check the reading of the adult subject who reads both languages. Method of Selecting Subjects Lists of all the boys who were eight, eleven, and fourteen years of age were made for three school in Austin. 6 Names of English-speaking children who showed exceptional advancement or very serious retardation were checked from the lists. Those who were left were given Form Vof the New Stanford Reading test, and again the very high and the very low were checked off. No cases of exceptional advancement were found among Spanish-speaking children, in fact, only one-Spanish-speaking boy was found with normal age-grade placement, so it was only necessary in their case to throw out the names of those who made very low scores. Those who were left after the elimnation based on the reading test were given the Thorndike-McCall Reading Test, Form W of the New Stanford Reading test, and three forms of the Burgess Picture Supplement Test. One hundred fifty-three English-speaking boys and fifty-one Spanish-speaking boys started the testing program. In the case of the accelerated group of English-speaking children the selection was made with a view to securing the youngest children that could be found who had the same ability as the Spanish-speaking children as measured by tests. The New Stanford and the Thorndike-McCall reading tests proved too difficult for the eight-year-old children so they were given Forms A and B of the Detroit Reading Test, Test I, and Types I, 11, and 111 of the Gates Primary Reading Test. •After the testing program the age and ability groups were paired according to mean reading ages expressed in months and according to mean raw scores on the Picture Supplement Test* The mean reading ages were computed from raw scores made on three forms of the test. Table I shows how closely the groups were paired with reference to reading ability as measured by the reading tests. ♦These standard deviations are based on 18 measures for the first two pairs and 36 measures for the last pair. They are given here merely to show the absolute variability in ability within the groups themselves. It may be seen that there is a wider range of ability on the eleven-year level than on any other. The English-speaking boys had a slight advantage over their corresponding age groups on all of the reading tests except in the case of eleven-year-old subjects. At this age the English-speaking group had 5.6 months higher reading age and 1.1 points lower Picture Supplement score. Since these differences were opposite in direction they were not considered sufficiently large to invalidate the pairing of these groups. The eight-year-old groups were paired on the basis of scores made on three forms of the Gates Primary Reading Test. The mean reading ages follow: English-speaking, 92.3 months; and Spanish-speaking, 92.4 months. Thus there is an insignificant difference in favor of the Spanish-speaking children. In order to get some measure of the standing of the different groups of subjects on material which did not depend entirely upon reading ability, two forms of the New Stanford Arithmetic Computation Test were given. A measure of mental ability was secured through use of the Pintner Non-Language Mental Tests. Table II shows the standing of the different groups on these tests. Although the different groups were closely paired with reference to reading age and to raw scores made on the Picture Supplement Test, some unstandardized measures taken from the tests show differences between them and agree with results that were later obtained with the eyemovement technique. It seemed desirable to get some measure of the rate of reading, the relative standing in comprehension, and some measure of the tendency of a subject to run carelessly over material within his range of difficulty without comprehending it. For a rate score the total number of paragraphs right, wrong, and omitted up to the last one tried on the two forms of the New Stanford Reading Test was used. The T-score on the Thorndike-Mc- Call Reading Test was taken as a measure of comprehension since the time limit was so great that all the subjects had ceased to work on it before time was called. The ’’carelessness score” was secured as follows: Since the Picture Supplement Test has all of its items of the same difficulty, a subject who can read one of them can read them all. If a paragraph is wrong or omitted it may be taken as an evidence of carelessness. The percent of the paragraphs wrong or omitted of the number right was taken as this measure. No paragraph was counted omitted after the last one tried. The standing of the groups is shown in Table 111. ♦The Thorndike-McCall T-score is a standardized score, but not in the particular sense in which it is used here. These unstandardized measures seem to indicate that the fourteen-year-old English-speaking subjects read faster, showed better comprehension and were more careful than the Spanish-speaking subjects of the same age. The reverse is true for the eleven-year-old groups except for carelessness. The youngest English-speaking group is between the two eleven-year-old groups except in comprehension. On this last measure they surpassed all but the fourteen-year-old English-speaking group. Material Read Before the Eye-Movement Camera The plan for taking the eye-movement records for all but the eight-year old boys was as follows: After the subjects had been paired they were taken before the eye-movement camera and given three practice selections to read. These selections were taken from forms of the tests used in pairing the groups, and were supposedly of increasing difficulty. At the end of each selection a blank was left for the subject to fill orally after he had read the selection silently. This oral response served as a signal that he had finished and furnished a test of comprehension. Thus the subject made the same kind of response before the camera that he had made during the period of testing except that here the response was oral instead of written as when he took the tests. Following the practice selections he was given four final selections for which eye-movement records were made. If the subject made an acceptable response when he reached the blank he was given credit for the selection; if not, the selection was not counted in the later study of results. This procedure is justified on the grounds that it is impossible to interpret a record of eye-movements made in reading a selection that is not comprehended. The final selections were chosen on this basis: Selection I was the well-known story of the ’’Lion and the Mouse” adapted from a story in a first reader. It was purposely chosen because it is so easy that any child who has made the most elementary reading adjustment can read it readily and comprehend it. Selection II was taken from the Stanford Achievement Test from the level of difficulty that is supposedly reached by eight-year-old children. Selections 111 and IV were taken from the same source and from the eleven-and fourteen-year levels respectively. Both the practice and the final selections are reproduced here for the benefit of the reader. Practice Selections in Order of Presentation Helen and Kate pulled their sled through the deep snow to the top of the hill and soon were coasting down again. They did this over and over. The snow was so deep that they found it hard work to drag the sled to the . Once a black raven wanted to have white feathers like a swan. The raven saw that the swan lived in the water, and thought that it was the water that made the swants feathers so white. So the raven decided to wash his feathers every day and see if it would make them . When the bear appeared near the hut, Walter was alone. His father had driven to the village, that morning, several miles away. Fortunately he had left his gun hanging on the wall loaded and ready for service. Walter was excited, but he did not hesitate. Quickly seizing the gun he shot the _. Final Selections A lion fell into a net one day while he was walking in the woods. He could not get away so he began to roar. A little mouse heard him and ran quickly to the lion. “Why are you roaring so, Mr. Lion?” asked the little mouse. “I am caught in the ropes and I cannot get away,” said the . Once there was a little girl who used to cry all day because she wanted the stars to play with. So one day she went out to find them. She walked until she was so tired she could go no farther. Just then a fairy appeared and asked where she was going. “I am going to hunt for the stars,” said the little girl, “but lam very tired.” Then the fairy reached up and pulled down a beautiful Age and Youth sat together by the fireside. Age, who was gray, bent, and infirm, talked incessantly of his childhood friends who were no more, of his successes and failures, and of the brevity and disappointments of human life. As Age spoke, Youth listened without saying a The planet Neptune is known to be exceedingly cold; far colder, in fact, than our north pole. The planet Mars, on the other hand, is blanketed with air and is probably almost as warm as the earth. So, although animal life may possibly exist upon Mars, it is unlikely that there is any life whatever upon the planet Four of the Spanish-speaking boys were able to read in their native language. Eye-movement records on two of the selections which had been translated into Spanish were taken. The selections are reproduced below: Un dia un leon que iba andando por un campo cayo en una red. No podia escaparse y empezo a rugir. Un ratoncito lo oyo y corrio pronto a donde estaba el leon. — Por que ruge asi Senor Leon?-*- pregunto el ratoncito. —Estoy enredado en los lazos y no puedo escaparmedijo el ♦ Una vez habia una ninita que lloraba todo el dia porque querida las estrellas para jugar con ellas. Un dia salio a buscarlas. Anduvo hasta que se canso tanto que ya no pudo andar mas adelante. En ese momento se aparecio un hada y le pregunto que adonde iba. —Voy a buscar las estrellas—contesto la ninita, —pero estoy muy cansada.—luego el hada levant o la mano y ba jo una hermosa Simple selections written in large type were prepared for the eight-year-old children. They were given two practice selections and one final selection. Only the final selection is reproduced here. John has his old ball. Tom has a one. Tom will bring his new ball and they will play with it. Which ball will the boys play with? Summary The problem has been stated as a laboratory study of the reading habits of Spanish-speaking children who entered school with little or no knowledge of English and who have been taught to read in English under the direct method of instruction. The eye-movement technique was chosen for use in the study. Groups of subjects were selected at ages eight, eleven, and fourteen. Control groups of English-speaking children of the same chronological ages and the same reading ages were selected. Because of the fact that the Spanish-speaking children are usually retarded in their school work, another control group of English-speaking subjects was chosen of the same mean reading age as the combined eleven- and fourteen-year old Spanish-speaking groups, but who were accelerated rather than retarded. The older boys read three practice and four final selections before the camera, and the eight-year-old boys read two practice selections and one final selection before the camera. J. G.: ”A Study of Bilingualism,” Pedagogical Seminary and Journal of Genetic Psychology, K: 2 Saer, D. J.: ’’The Inquiry into the Effect of Bilingualism upon the Intelligence of Young Children,” Journal of Experimental Pedagogy, 6: 233-240; 266-274, 3 Smith, F.: "Bilingualism and Mental Development," British Journal of Psychology, 12: 270-282, 1923. J.: Le developpement du language observe chez un enfant bilingue. dhampion, pp. 5 Pavlovitch, M.: Le langue enfantin. Acquisition de Serbe, et de Francais par un enfani Serbe. Paris: dhampTon? PpTWT * g Garretson, 0. K.: ”A Study of Causes of Retardation Among Mexican Children in a Small Public School System in Arizona.” Journal of Educational Psychology, 19* 31 - 40, 1928. Manuel, Herschel T.: The Education of Mexican and Span! sh-gpeaking Children in Texas. The Tunct for fie search in the Social Sciences, University of Texas, 1930, p. 36. Q Tireman, L. S.: “Reading in the Schools of New Mexico,” Elementary School Journal. 30: 621-626, 1930. Herschel T.: op* cit., pp. 30-31. l°lbid., p. 33. 11 Buswell, G. T.: A Laboratory Study of the Reading of Modern Foreign Languages, Publlcations oF~The American and Canadian Committees on Mo de rn Language s, MacMillan, 1927. xii and 100 pp. x 3 _. These schools were Fulmore and Palm, elementary schools, and John T. Allan High School, a junior high school. Group Age Mean Read- ing Age Dif- fer- ence Stan- dard Devia- tion* Mean Picture Supple- ment Score Dif- fer- ence Spanish 14 131.9 13.16 5.8 English 14 135.8 3.9 13.74 6.3 .5 Spanish 11 123.8 15.97 5.5 English 11 129.4 5.6 8.94 4.4 1.1 Spanish 11 & 14 127.9 15.22 5.7 English 9.7 129.5 1.6 12.04 6.3 .6 TABLE I MEAN READING AGES AND PICTURE SUPPLEMENT SCORES OF THE GROUPS USED Group Arithmetical Computation Age Pintner Mental Index Pintner Mental Age Spanish 14 137.5 37.0 11.4 English 14 140.8 28.8 9.6 Spani sh 11 128.5 45.6 10.2 English 11 126.0 44.3 9.2 Spanish 11 & 14 131.3 41.3 10.7 English 9.7 124.0 53.0 9.9 Spanish 8 91.5 39.5 7.2 English 8 88.6 31.0 6.1 TABLE II MEAN ARITHMETICAL COMPUTATION AGES, MENTAL INDICES, AND MENTAL AGES Group Rate Comprehension Carelessness Spanish 14 109.5 40.4 61.3 English 14 115.0 45.1 57.8 Spanish 11 105.5 37.6 151.8 English 11 97.1 34.3 44.3 English 9.7 102.1 42.0 53.5 TABLE 111 UNSTANDARDIZED MEASURES OF RATS, COMPREHENSION, AND "CARELESSNESS” * (Group Means) CHAPTER II THE USE OF THE EYE-MOVEMENT TECHNIQUE IN THE INVESTIGATION OF READING HABITS What an Eye-Movement Record Indicates The measurements which may be taken from an eye-movement record are: (1) the exact position of each pause and the number of pauses per line or for any given unit or material read; (2) the length of each pause in twentyfifths of a second; and (3) the number of regressive movements. Various other measures can be calculated from these. An analysis of a record from the standpoints of number, location, and sequence of pauses or regressive movements may yield important facts other than quantitative data, as will be shown later in the discussion. The eye-movement itself is simply the result of a muscular contraction. The eye-movement can take place without reading, but reading cannot take place without the eye-movement. As has said, the eye-movement is a H symp- tom” of the mental processes involved in reading. When a subject reads and comprehends what was read, a record of his eye-movements indicates how he went about getting the thought from the printed page. The average number of pauses made on any given unit of material indicates the normal recognition unit of the reader. The mature or expert reader may recognize several words at one fixation, the poor reader may recognize only one, and the beginner may spell the word and fixate every letter. The normal recognition unit for numbers or for a foreign language would be decidedly different from the normal recognition unit for easy material. But in any case the average number of fixations per line or unit of material would indicate the length of the normal recognition unit for the kind of material read. The duration of the pause indicates the speed with which the unit is recognized. Various factors may operate to affect the recognition time; for example, kind of material or familiarity with material. The average duration of fixations indicates the normal recognition time for the kind of material read if it is fairly homogeneous with respect to kind and difficulty. Regressive movements are an inverse measure of ’’the precision of recognition and regularity of progress across printed lines. Buswell mentions four kinds of regressive movements: “The most common regressive movement occurs at the beginning of a line where the return sweep of the eye has failed to carry the fixation back to the first word in the line and an additional regressive movement is required to make the initial part of the line clear. This type of regressive movement persists up to the more mature stages of reading. “A second type of regressive movement appears in the records of a number of mature readers who continually try to make as few fixations per line as possible. ”A third type of regressive movement is caused by lack of word knowledge . . . . ”A fourth type of regressive movement consists of a random oscillation of the eye with no apparent plan on the part of the reader. This type of behavior has been appropriately named a ’confusion period*. The reader fails to get a clear perception of the meaning and accordingly sets up a series of eye-fixations which move back and forth over the area, causing difficulty. H 15 The Analysis of an Eye-Movement Record Hate I shows a record of the eye-movements of a very immature reader, and Plate II shows the record of a mature reader. Plate I shows very strikingly the difficulties encountered by an immature reader while reading a selection in a language different from that of his home and social contacts. This subject was a Mexican boy fourteen years old with a reading age of 111 months, composite of scores made on three tests as explained in Chapter I. His arithmetic computation age was 130.5 months, - composite of scores made on two tests. When measured by the Pintner Non-Language Mental Test he secured rating of 14.5 years mental age and a mental index of 51. He was co-operative and seemed eager to comply with all the requests that were made of him. His spoken English was somewhat broken. It should be noted how the personification of "Youth” and "Age” seemed to bother this subject and how he laboriously spelled out unfamiliar words. That he was an intelligent reader is shown by the fact that he read each line better than the one just preceding it, and by the fact that words which gave him trouble at the beginning of the selection were read easily at the end. Despite his labored reading or "deciphering" he made a satisfactory response at the end of the selection. It should be added that this was the sixth selection he read before the camera, so he had had time to adjust to the conditions of the experiment. This record stand out by contrast with the record shown in Plate 11, a record made by a graduate student of Spanish-American extraction. A detailed summary of the two records is given below. The first and last lines were not counted in either record because it was feared that normal reading would not result while a subject was getting started or while he was filling the blank at the end of the selection. Subject ODSI4 showed three of the four types of regressive movements very clearly. On lines, 3,4, 5, and 6 the backward sweep of the eye from the end of the previous line was not long enough and one or more further regressive movements were necessary to carry the eye back to the place where the beginning of the line was clear. The eye-movements recorded when the words "talked”,"incessantly”, and "human” were fixated illustrate the type of regressive movement that results from word difficulty. On the word "human”, for example, the subject fixated every letter, but not in order. The type of regressive movement that results from a confusion period was well illustrated at the end of line 1. The last fifteen fixations had no apparent order The subject was evidently grasping for the meaning. This subject’s span of recognition was so small that he made about twice as many fixations per line as would normally be expected for a reader of his age, and his recognition period was also about twice as long as it should have been. He made about two or three times the normal number of regressive movements. In the light of information available concerning this boy’s general mental development it seems safe to conclude that his difficulties in reading were chiefly due to the language handicap. Records reproduced by 16 Buswell for English-speaking pupils reading a foreign language are similar in many respect to the records of this subject. This record "becomes even more interesting and the hypothesis that many of the reading difficulties are traceable to unfamiliarity with the language becomes even more tenable when Plate I is compared with Plate 111. The last mentioned record was made by Subject 17 CSSII. This subject was eleven years old and tested as follows: reading age, 110 months; mental index, 38; and arithmetical computation age, 105 months. He spoke English fluently and read easy material almost as well as an adult. He made an average of 9.5 fixations per line with an average duration of seven twenty-fifths of a second each. He required only two hundred sixtysix twenty-fifths seconds to read the entire selection, about one-third the time required by the subject of I. He made an average of two regressive movements per line, and half of these resulted from failure to make a return eye-sweep sufficiently long from the end of one line to make the beginning of the next line clear. Reading Habits and Scores on Reading Tests The standardized reading tests furnish objective and valuable measures of reading ability. These instruments cannot, however, provide the basis for the kind of analysis that is furnished by a photographic record of the eye-movements made during the process of reading. Plate I clearly shows that a reader may have poorer fundamental reading habits than his test record indicates, and Plate 111 shows as clearly that a reader may have better fundamental reading habits than his test record indicates. If either the testing or the eye-movement technique furnished all the data necessary for measuring reading ability or for diagnosing cases needing remedial treatment it would obviate the necessity for using the other technique at all. When both techniques have been used it can be shown how one supplements the other. A reading test can furnish a valid and reliable measure of the rate of reading; the analysis of an eye-movement record shows whether there is a balance between length of recognition unit and recognition time. A reading test would indicate that a pupil needed remedial treatment to improve his rate; the analysis of his eye-movement record would do much to indicate the apecific treatment needed. Some examples of the use of the eye-movement record for this purpose are given below. If he made about the same number of pauses per line with few regressive movements,he would likely need drill for the purpose of increasing his recognition span. If he habitually made too many fixations on individual words together with regressive movements, while otherwise his progress across the page was satisfactory, the need for general vocabulary drill would be indicated. If his record were satisfactory with reference to the number of fixations but an abnormally Jong fixation period was required, the need for drill to reduce his recognition time would be indicated. The Eye-Movement Record and Comprehension The eye-movement record does not show the presence or absence of comprehension, but if it is known whether or not a given selection was comprehended, valuable cues to the mental processes involved may be secured. This use of the eye-movement record may be seen from an analysis of the records shown in Plates IV and V. Plate IV is a reproduction of the record of a fourteen-year-old Mexican boy with a reading age of 109 months and Picture Supplement score of 2. A summary of his record for this selection would be little different from a similar summary for any of the other three selections which he read before the camera. Apparently he did not realize that he had failed to give the correct response at the end. Plate V is another record for Subject OSSII who read the first three selections in a way that approximates the performance of an adult reader. At the beginning he had trouble with the word, ’‘Neptune”, but after he left this word he proceeded normally until he reached the parenthetical phrase, “in fact.” By the time he reached the middle of the fourth line, he was completely lost. He floundered on to the end without catching the meaning. Neither of these subjects would have received credit for this paragraph on a standardized reading test, and here the analysis would have necessarily ended. The eyemovement record, however, shows the details of a striking difference in performance. It should be added that these two subjects differed in reading age by only one month, and that their eye-movement records also indicated, with the exception of the records for Selection IV, that they were about equal in reading ability. The measures that can be secured by photographing the movements of the eye of a subject while he is reading have been enumerated and described. Illustrations have been given of the use of eye-movement records in analyzing the reading habits of individuals. In the next chapter the results will be reported of the use of this technique in analyzing the reading habits of the groups described in Chapter I. 13 Buswell, G. T.: ’’Fundamental Reading Habits: A Study of Their Development,” Supplementary Education Monographs, No. 21, 1922. Chapter lor this' monograph gives an excellent discussion of the relation of eye-movements to mental processes in reading. 14 Ibid., p. 10 15 Ibid., pp. 35 - 36. 16 Buswell, G. T.: The Reading of Modern Foreign Languages, Macmillan, 19277" Plates xT"ana' m.-ppT 57 amt 17 The code for identification of subjects is to be read as follows: the number shows the age group, the letter ”A” is used for a number in the case of an adult subject, the letter just preceding the number is M S M or ”E" to indicate the language group, and the letters before this identify the particular subject. Plate I. Silent Reading by Subject ODSI4 of Selection 111. Plate 11. Silent Reading of Subject SSA of Selection 111 Plate 111. Silent Reading by Subject OSSII of Selection 111 Pl&te IV. Silent Reading by Subject BGSI4 of Selection IV. Plate V. Silent Reading by Subject CSSII of Selection IV Subject Measure Line 2 Line 3 Line 4 Line 5 Total Av. 0DS14 Fixations 21 21 17 13 72 18.00 SSA 6 8 11 10 35 8.75 CDS14 Regressions 6 5 6 3 20 5.00 SSA 0 2 2 3 7 1.75 CDS14 Total 226 : 198 170 137 731 182.75 SSA Time 38 50 59 54 201 50.25 CDS 14 Average 10.7 9.4 10.0 10.5 SSA Durst ion of Fixa- tions 6.3 6.2 5.3 5.4 TABLE IV SUMMARY OF THE RECORDS OF A MATURE AND AN IMMATURE READER CHAPTER III EXPERIMENTAL FINDINGS AND RESULTS Preliminary Statement It should be stated at the outset of this chapter that the present research is a laboratory study of the reading habits of children who learn to speak one language at home and have their school work in another language. Spanishspeaking children were chosen as subjects because they belong to the foreign language group that is most numerous in the Southwest. The present investigation is not a racial or a national study. However, English-speaking groups were used for controls. If the material used before the eyemovement camera had been standardized for all ages and reading abilities no control groups would have been needed. In lieu of suitably standardized material control groups were chosen of two kinds: (1) groups of the same age and reading ability, as measured by tests, and (2) a group with the same mean reading age as the combined eleven— and fourteen—yearold Spanish-speaking groups, but who averaged 3.3 years younger. Spanish-speaking children are, on the average, somewhat retarded in their school work. Subjects were selected for tae older control groups of practically the same degree of retardation as the Spanish-speaking children, but the younger control group was composed of children who tested well above the norms for their ages on reading tests. The method chosen for presenting the data is as follows: (1) the eye-movement records of the various Spanish-speaking groups are compared with reference to the several measures than can be secured through the use of the eye-movement technique; (2) comparisons are made with the records of English-speaking groups of the same reading age; (3) records are compared for subjects who read the same material in both English and Spanish; (4) comparisons are made with adult performance; and finally, (5) some individual comparisons are made. The last type of comparisons will be reserved for another chapter. Where possible, when any phase of the subject is treated all the data relating to it will be presented before proceeding to the next topic. This will facilitate the drawing of conclusions at the end of the study and make the development easier to follow. Number of Records Used for Each Selection In making the analyses which follow, the general rule was to exclude all records where the subjects failed to show a satisfactory degree of comprehension. Since the selections used before the camera were of increasing difficulty, all the subjects were not able to comprehend every selection. Table V shows the number in each group whose records were satisfactory in this respect. The numbers given in this table are the numbers used in all subsequent tables and discussions unless specifically stated otherwise. The Time Required to Read a Given Unit of Material Two of the important factors in reading are rate and comprehension. As was shown in Chapter 11, the eye-move- ment technique furnishes an exact and objective measure of the rate of reading and permits an analysis of the elements that enter into it. A gross measure of the rate of reading can be secured by finding the time for any given unit of material. This measure will be considered first and the elements into which total time can be analyzed will be discussed afterward. A convenient unit of material to consider in the study of total time is the time per line. The choice of this unit makes the records for all the selections comparable and shows the relative difficulty of the selections. The average time per line per subject for each of the four selections for all the groups except the eight-year-old subjects is shown in Table VI. The time per line varied from selection to selection with a general tendency to grow larger as the selections increased in difficulty. The dropping out of subjects as a result of failure in comprehension served in a measure to conceal this tendency in the table. The eleven-year-old Spanish-speaking group required less time per line than the fourteen-year-old group on all selections except the fourth, despite the fact that the older group had a higher reading age, a higher comprehension score, and made a better showing on vocabulary tests. The older Spanish-speaking subjects required less time per line for Selections I and IV than was required by their corresponding age groups of English-speaking children, but more time for Selections II and 111. On the eleven-year level the Spanish-speaking subjects required less time for all selections and the percentages of excess time required for the English-speaking group 18 were large enough to be significant. The accelerated English-speaking group made a better showing than the combined eleven- and fourteen-year-old Spanish-speaking groups on the last selection, but the number showing satisfactory comprehension on this selection was reduced to three. When the records of the Spanish-speaking subjects were compared with the records of three adult readers who read the same material, it was found that on the average they required about three times as long to read a line as the mature readers. Adult performance was approached by only two of the subjects and then only on Selection I. They required an average of thirty-eight and forty twenty-fifths seconds respectively for each line of this selection while the average time for the three adults was thirty twenty-fifths seconds. If partial comprehension is accepted it is possible to make comparison based on the records for twenty-three lines per subject, or one hundred thirty-eight lines for each group of six. This number is sufficiently large to justify treatment to determine the reliability of the differences in group performances. This treatment is shown in Table VII. From Table VII it is clearly seen that at age 11 the Spanish-speaking group required less time for reading than the English-speaking group of the same age. The difference is such as to amount to virtual certainty for the group tested. There is no significant difference on the fourteen-year level and when the combined eleven- and fourteenyear-old Spanish-speaking groups were compared with an accelerated group of English-speaking children of practically the same reading age there was no significant difference. It is to be noted, also, that the eleven-yearold Spanish-speaking subjects did significantly, if not reliably better than the fourteen-year old Spanish-speaking subjects. The difference was 1.48 times the sigma of the difference. Eleven-year-old Spanish-speaking subjects made a more favorable showing than any other age group. Interpretat ion. As has suggested, our instruments for mea- suring the achievement of Spanish-speaking children, standardized as they are on English-speaking groups, may not be entirely suited for the purpose. The Spanish-speaking child of eleven may have a general development of reading habits better than would be judged from a study of his test performances. Also, it may be significant to note that many Spanish-speaking boys are out of school by the time they are fourteen as a result of economic pressure, and consequently, it may be possible to select better readers on the eleven-year level. Evidence of the normal growth to be expected in reading habits between the ages of eleven and fourteen may be lost because boys who do well in school are most likely to find employment. Since the eight-year-old groups read only one selection and since there were only five subjects in each group it is possible to present individual records for them. Table VIII shows a summary of their records with respect to average time per line. Despite the fact that the English-speaking subjects were retarded and that they did not test as high on a non language mental test or on an arithmetical computation test, they required less time for reading the selection than the Spanish-speaking group. The time per line for the English-speaking subjects was 8.5 per cent less than that of the Spanish-speaking subjects, using the record of the Spanish-speaking subjects as the base. Average Number of Pauses Per Line Two of the principal elements which enter into total time are the length of the recognition unit and the recognition time. The average number of fixations per line is a measure of the length of the recognition unit. This measure will be considered next and treated in the same manner as average time per line. Table IX shows the records for the groups with reference to the average number of pauses per line. On Selection I the Spanish-speaking subjects made a more favorable showing than the retarded English-speaking groups and a poorer showing than the accelerated English-speaking group. On Selections II and 111 they did better than the English-speaking subjects except on the fourteen-year level. On Selection IV they made better records than English-speaking subjects their own ages, but compared unfavorably with accelerated English-speaking subjects. It should be remembered that the English-speaking subjects were reduced to one-fourth their original number on this selection, or from twelve to three, as a result of failure to comprehend what was read. Again it is to be noted, that on three selections out of four the eleven-year-old Spanish-speaking subjects made fewer pauses per line than the fourteen-year-old group. By reference to the original records it was found that the best individual performances were made by members of the eleven-year-old group. Both groups made about twice as many fixations per line as the adult group. Summaries of the individual records with reference to the average number of fixations per line for the eight-year-old subjects are shewn in Table X. The subjects of Table X are arranged in rank order of reading age. The English-speaking subjects averaged fewer fixations per line and made a more favorable showing, with one exception, when an individual was compared with an individual of the same rank in reading age. Byusing the record of the Spanish-speaking subjects as a base, it was found that English-speaking subjects required 14.2 per cent fewer fixations. Average Duration of Fixations Another element in total time is the time required for a subject to recognize the portion of material in his fixation area. The measure of the average recognition time is the average duration of a subject’s fixations. Group comparisons with reference to this measure are shown in Table XI. The Spanish-speaking subjects showed a quicker recognition time than the English-speaking subjects with the exception of Selection II on the fourteenyear level. Since Selection I was read with a satisfactory degree of comprehension by all the subjects, it is possible to treat this data to determine the reliability of the differences between the groups. This treatment is shown in Table XII. One may see from Table XII that the Spanish-speaking subjects showed a briefer recognition time on all age levels as compared with either the retarded or the accelerated English-speaking subjects. The differences amount to more than twice the standard deviation of the difference on all but the eight-year level and for the Spanish 14 as compared with the English 9.7. Spanish 11 made the most favorable showing of any age group but this group was decidedly below adult performance. In order to see if a chance grouping happened to favor the Spanish-speaking subjects in every case, some different groupings were tried. Eor example, the upper three on the eleven- and fourteen-year levels of the Spanish-speaking groups were paired with subjects chosen in a similar way from the English-speaking groups. The grouping gave a mean reading age of 137.66 months for the Spanishspeaking subjects as compared with a mean reading age of 141.38 months for the English-speaking group. The result was a difference of .49 and a standard deviation of the difference of .23, or expressed another way a difference 2.13 times the standard deviation of the difference favoring the Spanish-speaking group. The same Spanish-speaking group was compared with the six best readers of the accelerated English-speaking group and the difference in average duration of fixations was found to be 1.55 times the standard deviation of the difference. A g a further and final test, this Spanish-speaking group was compared with six of the best readers, as measured by reading age, that could be obtained by any grouping which was an English-speaking group with a reading age of 148 months, or an average of almost a year higher reading age. The difference in average duration of fixations was found to be 1.04 times the standard deviation of the difference. In other words the chances were about 85 in 100 that a Spanish-speaking group would show briefer fixation than an English-speaking group. Various other groupings of English-speaking children were tried and although it was possible to find groups that made fewer fixations per line and required less total time, no group could be formed which equaled the Spanishspeaking group composed of the three highest in reading age at ages eleven and fcrarteen. English-speaking groups could be found with a higher mean mental age, a higher reading age, and a higher arithmetical computation age, but they did not recognize their normal units of reading material for a fixation pause as quickly as the Spanishspeaking subjects. In all cases a sufficient degree of comprehension was demanded to fill the blank correctly at the end of the selection. A possible explanation of the relative standings of the English and Spanish-speaking groups may be found when their regressive movements are analyzed. This possibility will be taken up below. The Number of Regressive Movements Another factor which affects speed of reading is the number of regressive movements, or backward sweeps of the eye* Four different kinds of regressive movements were defined and discussed in Chapter ll* The regressive move*- ment serves to slow down the reading process, and as a general rule poor readers make more regressive movements than good readers. However, as was pointed out in the earlier discussion, the good reader may endeavor to take as large units as possible at each fixation and as a result find it necessary to make some regressive movements in order to comprehend what he reads. Buswell2o has pointed out that more regressive movements may be made while the subject is making a transition from a lower to a higher level of reading habits. The gain from the larger fixation unit may more than offset the loss due to regressive movements. The average number of regressive movements per line made by the different groups for each of the four selections is shown in Table XIII. The Spanish-speaking subjects showed a clear advantage on all but the fourth selection. On this selection they surpassed their corresponding age groups, but the combined eleven- and fourteen- year-old Spanish-speaking group was surpassed by the accelerated English-speaking group. Again it should be remembered that the number of English-speaking children in this group who comprehended Selection IV was only three. The eleven-year-old Spanish-speaking subjects made fewer regressive movements than the fourteen-year-old group on the first three selections. All the groups made two or three times as many regressive movements as the adult ‘group. The number of regressive movements made by the eight-year-old subjects is shown in Table XIV. It is to be remembered that subjects are arranged in rank order of reading age. The English-speaking subjects had the advantage from the standpoint of the total number of regressive movements and from individual to individual according to rank in reading age, with one exception. Their records show 7.5 per cent fewer regressive movements. Kinds of Regressive Movements Made a the Different Age and Language Groups It is not always possible to look at an eye-movement record and classify all the regressive movements into the four different kinds of regressive movements discussed above. It is possible, however, to tell when the regressive movements result from failure to make the return eye-sweep from the end of the previous line sufficiently long to make the beginning of the line clear. In Table XV the total number of regressive movements, the number resulting from failure to make a sufficiently long return sweep of the eye from the end of the previous line, and the per cent this number is of the total are shown for all groups and selections. Most of the regressive movements made by two of the adults were of this type, but one adult subject made a number of regressive movements of the type that result from the effort to make too few fixations. In order to give a complete picture of the reading habits of all the subjects and on reading material of increasing difficulty the records of all the subjects were counted whether complete comprehension was shown or not. Two facts stand out from a study of Table XV: (1) the English-speaking groups made more regressive movements when considered from the standpoint of totals or when compared from selection to selection except on the eight-year level, (2) Spanish-speaking subjects tend to have a larger per cent of their regressive movements of the type that results from insufficient backward eyesweep from the end of one line to the beginning of the next. As yet the evidence is inconclusive as to whether this is to be taken as a sign of immaturity in reading. The Effect of Regressive Movements Upon Average Duration of Fixations Regressive movements tend to increase the number of fixations per line and thus increase total time. There is the possibility that regressive movements tend to increase the average duration of fixations and increase total time in this way also. By taking the length of each fixation just before a regressive movement and the length of each regressive fixation on Selection I for all the groups it was possible to collect some evidence on this point. After carrying this process through it was found that regressive movements do increase the average duration of fixations. The relative lengths of pre-regressive fixations and regressive fixations are shown in Table XVI. After studying the data presented in Table XVI, it seems perfectly reasonable to suggest that the more favorable showing of the Spanish-speaking groups above the age of eight with reference to average duration of fixations may be partly if not wholly explained by the fact that the English-speaking subjects make more regressive movements. Variability in Reading Performance All readers show some variability in reading performance. No two lines of any selection are likely to be of exactly the same difficulty and different combinations of words will be broken into different lengths of recognition units. Some readers show little variation from their mean performance on any specific measure. Such a reader might be mature or immature, which is to say that little variation from excellent performance would be indicative of maturity in reading and little variation from poor performance would be indicative of poor reading. But as the mature reader gets material that approaches his level of difficulty he will necessarily show some variability. However, the working out of the relationships between variability and various phases of reading performance is left for other researches. In this study the only question which is raised with reference to variability in performance is: do Spanish-speaking subjects tend to be more or less variable than English-speaking subjects of the same reading age, and are they more or less variable than adults? Mean coefficients of variability and group coefficients of variability are shown in Table XVII. The mean coefficients of variability are simply the means of the individual coefficients. The group coefficients were computed from group standard deviations and means. ♦Figured for Selection I only. In comparing any two groups it should be kept in mind that a lower mean coefficient of variability for a group signifies that on the average the individual members tended to show less deviation from their individual mean performances; a lower group coefficient signifies less deviation from the group mean. With reference to mean coefficients of variability the Spanish-speaking groups showed less variability than the English-speaking groups with the exception of: (1) total time and average duration of fixations on the fourteen-year level, and (2) average duration of fixations when both groups are compared with the accelerated group. All group coefficients for Spanish-speaking subjects were smaller than those of the accelerated English-speaking group, and all group coefficients were smaller for average duration of fixations, otherwise, the group coefficients of the Spanish-speaking groups were larger. The coefficients of variability for the adults were all smaller than those of the Spanish-speaking children. On the eight-year level the English-speaking children were more variable. On the average, the Spanish-speaking group showed a tendency to be less variable in reading performance than English-speaking groups. Out of eighteen comparisons they showed less variability with reference to eleven measures. Oomparison of Performance in Reading Parallel Selections in English and Spanish Four of the Spanish-speaking subjects were able to read easjr Spanish. Two of the selections read before the camera were translated into Spanish and read in both languages by these subjects. Probably because they had had much more practice in reading English than Spanish they read better in the former language. As can readily be seen from Table XVII they made more fixations, took a longer time per line, and made more regressive movements in Spanish than in English. However, the average duration of fixations was less for Spanish. With reference to average duration of fixations these subjects approached the record of the Spanish-speaking adult. His record was 6.08 twenty-fifths seconds for the first selection and 5.85 for the second. Two of the younger subjects made snorter fixations on the first selec< tion and one made a more favorable showing on the second. When the average duration of fixations in English and Spanish for this group were treated for reliability it was found that the chances were 94 in 100 that a fixation for reading Spanish would be shorter than a fixation for reading English. It is also interesting to note in this con- nection that more variability in the duration of fixations was shown for Spanish than English. Summary Material has been presented for eight groups who read one to six final selections before the eye-movement camera. All of the common measures secured by this technique for these groups have been presented and discussed. Such a variety of material needs to be summarized so the reader can get the main facts before him at one time. Eight-year-old Spanish-speaking children do not make as good a showing on any of the measures as English-speaking children of the same age, except on average duration of fixations. With respect to this measure it was found that the chances were 85 in 100 that a fixation by a Spanish-speaking child would be briefer than one by an Englishspeaking child. The Spanish-speaking children made a better showing on a non-language mental test in arithmetical computation. It would seem that their principal difficulty was in the use of the English language. Detailed individual comparisons will be made in Chapter IV. With respect to the reading of two selections in both English and Spanish by four of the subjects it was found that English was the favored language for reading except that shorter fixations were made in Spanish. Satisfactory comprehension was shown in every case. A summary of the comparative performances for the groups above the age of eight may well be given in tabular form. This material is presented, in Table XIX. The five groups compared are ranked in order of excellence on each measure. When ranks are added in the extension columns the lowest total signifies the best general record. ♦Ranks in reading age are not included in the totals. The grand total of the ranks shows the following totals arranged in increasing magnitude: Spanish 11 was first with a total of 33, English 9.7 second with 40.5, closely followed by Spanish 14 with 41, English 14 fourth with 51.5, and English 11 in last place with 75. The maximum total was 80 and the minimum 30. Spanish 11 led all the groups on the first three selections and made the minimum total for Selection 11. On Selection IV they were in third place. Spanish-speaking subjects on the fourteen-year level surpassed the English-subjects of both the eleven- and fourteen-year levels and almost equalled the accelerated group. It is to be noted that ranks in reading age were constant with the exception of the exchange of places by Spanish 14 and English 9.7 on the second selection. With reference to average duration of fixations and average number of regressive movements per line the Spanish-speaking groups showed a clear advantage except for the eight-year level, where they made more regressive movements. Where group performances were considered, the read, ing habits revealed by the eye-movement technique did not indicate growth from age eleven to age fourteen. Some of the best individual records, however, were made by subjects on the fourteen-year level. On none of the measures except average duration of fixations did the Spanish-speaking subjects approach adult performance. A few individual readers, however, made records that compared favorably with adult records. Individual records will be presented and comparisons made in Chapter IV. In addition some analyses will be made that have not received statistical treatment. 1 8 See Table VII, below, for reliability of differences H. T., op. cit., p. 36. 20 Buswell, G. T.: Fundamental Reading Habits: A Study of Development, Supplementary Uducat ion Monographs 7 No. 31, p. 36• Group I Selection IV II III Spani sh 14 6 6 6 3 English 14 6 6 6 4 Spanish 11 6 5 5 4 English 11 6 5 6 2 Spanish 11 & 14 12 11 11 7 English 9.7 12 9 11 3 TABLE V NUMBER OF SUBJECTS COMPREHENDING EACH SELECTION Selection Part I Spani s^° U fngli sb Differ- ence Per cent of excess of English over Spanish Age 14 Age 14 I 68.7 77.9 9.2 13.39 II 73.5 72.9 -0.6 -8.16 III 81.0 77.0 -4.0 -4.93 IV 71.6 86.5 4.9 6.84 Part II Age 11 Age 11 I 64.0 79.2 15.2 23.75 II 59.3 81.9 22.6 38.11 III 66.9 105.3 38.4 57.39 IV 90.9 122.6 31.7 34.87 Part III Age 11 and 14 Age 9.7 I 66.3 67.6 1.3 1.96 II 67.0 68.7 1.7 2.53 III 74.6 80.0 3.4 4.55 IV 81.3 58.7 •22.6 -27.79 TABLE VI AVERAGE TIME PER LINE Group Mean time per line S.D. S.D. of Mean Differ- ence (E - S) S.D. of differ- ence D ‘ S.UTof differ- ence Spanish 14 77.33 26.93 2.29 .03 2.90 .01 English 14 77.36 20.13 1.71 Spanish 11 English 11 72.06 96.37 32.06 36.58 2.73 3.11 24.21 4.14 5.87 Spanish 11 & 14 English 9.7 74.70 79.05 28.86 47.48 1.73 4.04 3.41 4.39 .77 TABLE VII THE RELIABILITY OF THE DIFFERENCES BETWEEN THE GROUPS FOR AVERAGE TIME PER LINE ON FOUR SELECTIONS Spanish- speaking subjects Average Time per line English speaking subjects Average Time per Line EC 92.75 MF 108.25 FG 129.00 MO 162.50 FR 139.25 RB 77.50 TG 83.25 WB 67.75 AA 146.00 JB 123.50 Total time for all subjects 2361.00 2158.00 Average time per subject 472.20 431.60 Average time per subject per line 118.05 107.90 TABLE VIII AVERAGE TIME PER LINE FOR EIGET-YEAR-OLD SUBJECTS Selection Part I Per cent of excess of English over Spanish Group Spanish English Differ- ence Age 14 Age 14 I 10.4 11.0 .6 5.76 II 10.6 10.3 -.3 -2.83 III 10.7 9.6 -1.1 -10.28 IV 11.0 11.6 .6 5.45 Part II Age 11 Age 11 I 10.0 11.3 1.3 13.00 II 9.3 11.9 2.6 27.95 III 9.7 13.1 3.4 35.05 IV 12.7 15.1 2.7 21.77 Part III Age 11 & 14 Age 9.7 I 10.2 9.7 -.5 -4.90 II 9.7 9.8 • 1 1.03 III 10.2 10.5 .3 2.94 IV 11.9 8.5 -3 t 3 -27.73 TABLE IX AVERAGE NUMBER OF PAUSES PER LINE . Average Number Spanish o f Fixations English Average Number of Fixations Per Line per line EC 13.50 MF 11.50 FG 13.25 MO 15.25 FR 14.75 RB 10.50 TG 7.75 WB 7.00 AA 16.00 JB 12.00 Total Fixations 261.00 225.00 Average Number 52.50 45.00 Fixations per subject Average Number 13.16 11.25 Fixations per subject per line TABLE X AVERAGE NUMBER OF FIXATIONS PER LINE (Eight-year-old Subjects) Part I Group Per cent of excess ■ of English over Spanish Selection Spanish English Age 14 Differ- ence Age 14 I 6.57 7.05 .48 7.30 II 10.54 7.01 -3.53 -33.49 III 7.56 7.74 .18 2.38 IV 6.19 7.45 1.26 20.35 Part II Age 11 Age 11 I 6.37 6.90 .53 8.32 II 6.33 6.87 .54 8.53 III 6.90 8.00 1.10 15.94 IV 7.24 8.11 .87 12.01 Part III Age 11 & 14 Age 9.7 I 6.47 6.74 .27 4.17 II 6.66 7.03 .37 5.55 III 7.28 7.57 .29 3.98 IV 6.82 6.90 .08 1.17 TABLE XI AVERAGE DURATION OF FIXATIONS Group Average Duration S.D. Fixation S.D. of Mean Dif- fer- ence S.D.of differ- ence D Diff. Spanish 14 6.54 2.53 .159 English 14 7.11 2.91 .179 .57 .24 2.37 Spanish 11 6.37 2.56 .164 English 11 6.96 3.24 .196 .59 .25 2.36 English 9.7 6.83 2.94 .134 Spanish 11 (Compared with .46 .21 2.19 Spanish 14 ( English 9.7 .29 .208 1.39 Spanish 11 (Compared with Sp.14) .17 .22 .77 Spanish 8 9.04 5.3 .328 English 8 9.57 6.0 .400 .54 .51 1.05 Adult 5.48 1.97 .240 Spanish 11 (Compared with Sp. - .95 .29 3.27 adult) TABLE XII RELIABILITIES OF GROUP DIFFERENCES IN AVERAGE DURATION OF FIXATIONS Selection Part I Per cent of excess of English over Spanish Group Differ- ence Spani sh Age 14 English Age 14 I 1.66 2.41 .78 46.98 II 1.83 2.16 .33 18.03 III 1.37 1.79 .42 30.65 IV 1.75 2.50 .75 42.85 Part II Age 11 Age 11 I 2.00 2.54 .54 27.00 II 1.63 2.90 .27 16.56 III 1.75 2.82 .07 4.00 IV 2.85 3.10 .25 8.77 Part III Age 11 & 14 Age 9.7 I 1.83 2.39 .56 30.60 II 1.74 2.16 .42 24.13 III 1.54 2.13 .59 38.31 IV 2.51 1.65 -.86 -34.26 TABLE XIII AVERAGE NUMBER OF REGRESSIVE MOVEMENTS PER LINE Spanish mber Regressive Movements English Number Regressive Movements EC 17 MF 14 FG 14 MC 18 FR 11 RB 10 TG 5 WB 4 AA 19 JB 15 Total Regressive Movement s 66.0 61.0 Average Number per Subject 13*2 12.5 Average per subject per line 3.3 3.1 TABLE XIV REGRESSIVE MOVEMENTS OF THE EIGHT-YEAR-OLD GROUPS Selec- tion Total Regres sive Move- ments Spanish Age 14 - Fai- lure to can back Per Cent -y Total Regres- sive Move- ments English Age 14 Failure to carry- back Per Cent I 40 15 37.50 58 16 27.59 II 44 18 40.91 78 25 32.05 III 33 10 30.30 43 9 20.93 IV 63 21 33.30 67 19 28.36 Total 180 164 35.50 246 69 28.00 Age 11 Age 11 I 48 14 27.17 61 21 34.43 II 71 29 40.85 102 28 27.45 III 41 11 26.80 67 17 25.37 IV 79 18 22.78 96 27 28.13 Total 239 72 24.20 326 93 28.50 Age 11 and 14 Age 9.7 I 88 29 33.95 115 36. 31.30 II 115 47 40.80 177 46 25.90 III 74 21 28.38 94 22 23.40 IV 142 39 27.40 153 33 21.59 Total 419 138 32.90 539 137 25.40 Age 8 Age 8 Special . 66 9 13.60 61 15 24.60 TABLE XV NUMBER OF REGRESSIVE MOVEMENTS RESULTING FROM FAILURE OF RETURN EYE-SWEEP FROM END OF ONE LINE TO CARRY BACK TO BEGINNING OF THE LINE Group Number Regressive Movements Total Time Pre-re gressive Total Time Regre sive D. s- Per Cent Excess of Reg, Av.Dura- tion Pre- Regres- , sive Av. Dura- tion Reg. Spanish 14 36 214 223 9 4.2 5.94 6.19 English 14 46 290 333 43 14.8 6.30 7.23 Spani sh 11 39 229 275 46 20.0 5.87 7.05 English 11 51 298 385 87 29.1 5.84 7.54 Spanish 11&14 75 443 498 55 12.4 5.90 6.64 English 9.7 98 620 731 111 17.9 6.32 7.45 TABLE XVI LENGTH OF PRE-REGRESSIVE FIXATIONS COMPARED WITH REGRESSIVE FIXATIONS Group Mean V Total Time Group V Mean V Total Fixation Time Per line Group V Fixation Per line Mean V* Group V Average Average Duration Dura. Spanish 14 English 14 28.10 24.30 34.82 26.02 21.12 22.88 28.18 26.35 38.41 35.42 38.68 40.92 Spanish 11 English 11 26.20 33.66 44.49 37.95 24.83 27.24 36.36 31.16 37.92 41.99 40.18 46.55 Span. 11&14 English 9.7 27.15 35.97 38.66 55.67 21.97 31.71 32.29 40.44 38.17 36.03 39.31 43.04 Adult 19.80 28.39 20.66 26.58 31.63 36.34 Spaish 8 English 8 55.55 63.43 58.62 62.56 TABLE XVII VARIABILITY IN EYE-MOVEMENT PERFORMANCE Part I Average Time Per Line Average Number of Fixations English Spanish English Spanish AC 63.25 74.20 8.75 13.20 58.33 79.14 8.66 11.71 CP 55.60 77.20 9.00 11.60 54.00 60.14 8.16 10.00 FE 77.50 111.80 11.50 16.60 89.33 89.14 13.00 14.50 PC 61.75 66.20 9.50 11.60 59.33 52.00 9.33 9.14 Group 64.56 82.35 9.68 13.25 65.28 69.14 9.79 11.35 TABLE XVIII COMPARISON OF READING PERFORMANCE IN SPANISH AND ENGLISH TABLE XVIII (continued) Part II Average Length of Fixation English Spanish Average Number of Regressive Movements English Spanish AC 7.22 6.75 5.60 6.75 1.00 1.00 2.20 2.42 OP 6.19 6.61 6.65 6.01 1.75 1.00 2.20 2.00 FE 6.73 6.87 6.70 6.27 1.75 3.83 3.80 4.28 PC 6.50 6.35 5.65 5.68 .50 .50 1.60 .85 Group 6.66 6.66 6.81 6.08 1.37 1.33 3.45 3.39 Group Read- ing Age Avg. Num- ber of Fix. Average - Number Regres- sive Move- ments Per line Aver- age Time Per Line Average Duration of Fixation Total* Selection I Spanish 14 2 3 1 3 2 9 English 14 1 4 4 4 5 17 Spanish 11 5 2 2 1 1 6 English 11 4 5 5 5 4 19 English 9.7 3 1 3 2 3 9 Selection II Spanish 14 3 3 2 4 5 14 English 14 1 4 3.5 3 3 13.5 Spanish 11 5 1 1 1 1 4 English 11 4 5 5 2 17 English 9.7 2 3.5 § 4 11.5 Selection III Spanish 14 2 4 1 4 2 11 English 14 1 1 3 2 4 10 Spanish 11 5 2 2 1 1 6 English 11 4 5 5 5 5 so English 9.7 3 3 4 3 3 13 Selection IV Spanish 14 2 2 2 2 1 7 English 14 1 1 3 3 4 11 Spanish 11 5 5 4 4 3 16 English 11 4 4 5 5 5 19 English 9.7 3 3 1 1 2 7 TABLE XIX RANKS OF THE GROUPS WITH REFERENCE TO THE VARIOUS MEASURES CHAPTER IV INDIVIDUAL ANALYSES AND COMPARISONS In a study of this type much valuable material would be lost if the data were given group treatment only. Some of the most valuable evidence upon which to base conclusions with reference to the reading habits of Spanishspeaking subjects have been found from the analyses of individual records. In the present chapter, therefore, the analyses will be in terms of individual eye-movement records. In some cases records will be reproduced for each individual considered and in other cases a single record may serve to illustrate eye-movement habits common to several subjects. The latter plan will be followed whenever it seems practical. Individual Analysis of the Reading of An Eight-Year-Old Spanish-Speaking Boy MM* 3MMMMKBB aMMBBROW IHHBiBMMaMMKMMEaMEM* 4fIOhBiMUMMNkMKSKNB iMMBMOM* The first record chosen for analysis is that of subject ECSB, an eight-year-old Spanish-speaking boy with a reading age of 103.2 months as measured from a mean score on three forms of the Gates Primary Test. He was in grade 3a at the time he was tested. He had a mental index of 59, nine points above normal, on the Pintner Non-Language Tests, and an arithmetical computation age of 107.5 months, a mean score for two forms of the New Stanford Arithmetic Computation Test. His record is reproduced in Plate VI. It will be noticed that this subject fixated practically every letter of every word he read. In some cases there were more fixations per line than there were letters. In a number of instances he fixated the last part of a word and had to make one or two regressive movements to make the beginning of the word clear or to get the meaning. About one third of his fixations were regressive fixations. He required 371 twenty-fifths seconds to read the selection, made 54 fixations, and 17 regressive movements, not counting the first and last lines. His averages per line were 92.25 twenty-fifths seconds average time, 13.5 fixations, 4.25 regressive movements, and his average fixation time was 6.87 twenty-fifths of a second. This subject was far superior to any one on his age level from the standpoint of achievement tests and second only to one subject with reference to his eye-movement habits. From a study of this record it is evident that only the most elementary reading habits have been formed. Still it may be taken as fairly typical of the records of these children who are in the early stages of learning a second language. Subject ECSB was chosen from the highest position in his group from the standpoint of his test performances. In order to get a comparison of his record with the record of a subject who did not have the language handicap, his record should be compared with the record made by an English-speaking subject who ranked first in his group. This subject, MFEB, was in grade 2a, had a reading age of 97.4 months, a mental index of 33, seventeen points below normal, and an arithmetical computation age of 96 months as measured by the test described above. He tested below the Spanish-speaking subject in every respect. His record is reproduced in Plate VII. A longer total time for reading the selection was required by Subject MFEB than was required by the Spanish-speaking subject with whom he was compared, but on the whole a more mature type of reading habits was shown by the English-speaking subject. He made eight fewer fixations on the whole selection and an average of two fixations less per line. He made fewer regressive movements. When his record is studied closely it is found that he had three distinct “confusion periods, 0 but aside from these he read along in a fairly rhythmical fashion. More of his regressive movements were of the type that result from failure to make a return eye-G sweep sufficiently long from the end of one line to make the beginning of the next line clear. Two records for the English-speaking subjects in this group were superior in several respects to the record of MFEB. The two records presented here may constitute insufficient evidence upon which to base the conclusion that the language handicap of the Spanishspeaking child is shown by his difficulty in developing desirable reading habits. A study of all their records however, does seem to indicate that this is an important factor. The records just shown in Plates VI and VII repre- sent the best reading found on the eight-year level as indicated by tests. These records should be contrasted with the two best records as indicated by eye-movement records. The first two subjects had not made progress beyond what has been called the ”deciphering” stage in reading, while the two subjects whose records are presented in Plates VIII and IX had progressed beyond this stage even if their test records do not indicate it. It should be stated that all the subjects whose records were counted on the eight-year level answered the question at the end of the selection they read, and other questions on the ”story”. The subject whose record is shown in Plate VIII, TGSB, had a reading age of 88 months and the subject whose record is shown in Plate IX, WBEB, had a reading age of two-tenths of a month less. The two subjects were well paired so far as the reading tests show. A study of these records shows that both of these subjects read this simple selection in a straight forward and fairly rhythmical fashion. The English-speaking child was, however, superior in every respect as is seen from the summary of the two records given in Table XX. Growth in Eye-Movement Habits for Spanish- Speaking Subjects Mention has already been made of the fact that growth in reading habits is not shown between the ages eleven and fourteen for Spanish-speaking children when they are considered in groups. In fact, the eleven-year-old subjects showed more mature eye-movement habits, on the average, than the older group. However, where small groups are considered it is entirely possible for one or two records of outstanding excellence or inferiority to magnify or conceal group tendencies. So it seems desirable to make some individual comparisons before stating a final conclusion with reference to growth in eye-movement habits from one age level to the other. It is possible to make individual comparisons, where reading age is held constant, since three members of the eleven-year-old group had almost exactly the same reading ages, individual to individual, as three members of the fourteen-year-old group. The test records for these subjects are shown by pairs in Table XXL The figures of the identification codes for the subjects,as was stated in Chapter 11, indicate their ages. *The smaller score is more desirable in this case. It may be seen that the scores for the individuals run very close together in almost every one of the test measures. The plan followed in reproducing records for individual comparisons was to choose the record for the most difficult selection read with satisfactory comprehension by both members of the pair. The records for Selection IV are reproduced for the pair with the highest reading ages. Plates X and XI show these records. Both of these records show a fairly mature type of reading habits. There are no marked ’’confusion periods” shown for either sugject. Both subjects required fifteen twenty-fifths of a second for the word ’’Neptune” and both subjects made a regressive movement on this word. The older subject was more rhythmical in his reading as was evidence by the fact that he made fewer regres' sive movements. Not counting the first and last lines the older subject made three regressive movements for the entire selection, two of these were of the most common type, involving failure to carry the eye back far enough from the end of one line to make the beginning of the next line clear. His other regressive movement seems to have resulted from an effort to include too many words in his recognition span when he reached the expression, ”our north pole”. At this point he went back to the word ”north” for a very brief pause. This seems to be a logical thing to do since either ’’north” or ’’south” could have been used here without altering the congruity of the sentence. The younger subject made six regressive movements, two of which were of the most common type as explained above. Three of his other four regressive movements were made in connection with ”relationship” words, “than”, ”with” and ”so”, and one seemingly resulted from his trying to take too long a fixation unit. Summaries of their records will also be given in Table XXII with those of the other subjects considered in this connection. After a study of these records, it may be stated that on the higher level of reading ability the records for a pair of subjects with approximately the same reading ages seem to indicate better eye-movement habits for the older subject. Plates XII and XIII show the records of PCSI4 and CPSII. These records show reading habits similar to those that would be shown by normal readers in the fourth or fifth grade. The older subject required more time and made more fixations, but he required less time for each fixation and made only one-half as many regressive movements. The older subject filled the blank at the end of the selection much more readily than the younger subject. It is difficult to choose the better of these records, but in some respects the record of the older subject shows greater maturity of reading habits. Summaries of the eye-movement records of these subjects will also be given in Table XXII. The next pair of records, Plates XIV and XV, show a distinct advantage in favor of the younger subject. These subjects, SHSI4 and TOSH, have exactly the same reading age and are otherwise fairly well-matched, but the younger subject made fewer and briefer pauses and fewer regressive movements. The older subject had much more difficulty with the personifications ’’Age” and M Youth H . Summaries of their records are also shown in Table XXII. Growth in eye-movement habits between the ages of eleven and fourteen is not clearly shown for Spanishspeaking children when they are considered in groups or as individuals when reeding age is held constant. One exception, however, is found in the case of regressive movements. The older subjects made fewer regressive movements. In the case of English-speaking subjects it was found that where the readers had formed reasonably mature eye-movement habits the younger of two readers of the same reading age invariably read better. On the other hand, it was possible to find some subjects on the eightand nine-year age levels who had formed very immature eye-movement habits, but who have reading ages as high as normal eleven-year-old- children. Such subjects are far below the average on all the eye-movement measures. There is some slight evidence of growth with age for Spanish-speaking children even after fairly mature eye-movement habits have been formed. The only measure that snows this growth conclusively is the measure of the regressive movements. The record of the younger of a pair of English-speaking children of the same reading age but on different age levels is usually better than that of the older if the reader has progressed past the ’’deciphering” stage that is characteristic of the beginner. Young readers in either group may test as high as subjects who show marked superiority in eye-movement habits. Comparison of the Record of an Eleven-Year-Old Span ish-Speaking Subject with the Records of Five English-Speaking Subjects With Approximately the Same Reading Age Summaries of the eye-movement records of five English-speaking subjects and one Spanish-speaking subject are shown in Table XXIII. All of the subjects had reading ages that varied less than four months from 128, the reading age of the Spanish-speaking subject. Such a comparison should furnish some evidence as to whether the Spanish-speaking subject has as mature reading habits for his reading age as indicated by reading tests as the English-speaking child. The summaries of the eye-movement records for six subjects contained in Table XXlllindicate that the Spanish-speaking subject, ACSII, has as good or better fundamental reading habits for his reading age as the English-speaking subjects. Only one subject, CANES.B, showed more mature reading habits than subject ACSII. Their records for both selections showed the same number of regressive movements per line. All the subjects except ESEB.3 had formed eyemovement habits that place them above the ”deciphering" class. The eye-movement records of the subjects whose records are summarized in Table XXIIIare reproduced in Plates XVI, XVII, XVIII, XIX, XX, and XXI. Difficulties of Spanish-speaking Children With Figures of Speech, Unusual Words, Etc« It has been remarked from time to time when individual records were being considered that the subject had difficulty with the personification of ’’Age” or “Youth” or with some unusual word. When the records of immature Spanish-speaking subjects were analyzed, it was found that they had more diffi culties of this sort than English-speaking children of the same reading age. The record of CDSI4 which was reproduced in Chapter II illustrates this kind of difficulty. Several records of subjects who had reading ages below those used in the group comparisons were studied and were found to be similar to the record of CDSI4 in this respect. This finding raised the question as to whether a study of all the records of Spanish-speaking subjects would indicate a tendency toward greater difficulty on figures of speech and unusual words than would be found for English-speaking subjects. In order to answer the question the number and duration of fixations for all the subjects were found for the words ’’Age”, ’’Youth”, and ’’planet”. The results are shown in Table XXIV.\ From this table it may be seen that, on the average, the Spanish-speaking subjects did as well or better than the English-speaking subjects of the same age and reading ability and better than the accelerated English-speaking group on the personifications of “Age” and “Youth”. The English-speaking subjects made a better showing, however, on the word “planet”. Comparison of Groups with Reference Eye-Movement Records for the Last Line of Selection I. When a subject finished reading a selection before the eye-movement camera, he was required to supply a word which completed the meaning of the last sentence of the selection. A satisfactory response was taken as evidence that the subject had comprehended what he read. This kind of comprehension test was chosen, as was explained in Chapter I, because the form of response required was identical with that required in the reading tests used in selecting the experimental groups, accept that an oral response was given before the camera where a written response was made on the reading tests. This further difference should also be pointed out: on the reading tests the subject could read any given selection as often as was necessary in order to make a satisfactory response, while his instructions before the eye-movement camera were, “Read the selection one time and speak the word which fills the blank at the end.” Thus, in order to receive credit for a selection a subject had to comprehend it at one reading. Measures of the readiness with which satisfactory responses were given may indicate something of the maturity of the readers and should be particularly valuable in a study of bilingual subjects. Such measures are available from analyses of the eyemovement records. Plates XXII, XXIII, XXIV, and XXV illustrate differences in the readiness of response when the blank at the end of a selection is reached* In Table XXIV the average number of fixations and regressive movements per subject, and the average duration of fixations for the last line of Selection I are given. This group summary will furnish the basis for evaluating individual performances. On the fourteen-year level, the Spanish-speaking group made slightly fewer fixations and regressive movements than the English-speaking group, but their average duration of fixations was longer. The eleven-year-old Spanish-speaking subjects were surpassed by their corresponding age group of English-speaking subjects, except that their average duration of fixations was slightly shorter. The accelerated English-speaking groups made a more favorable showing on all the measures than either group of Spanish-speaking subjects except with respect to average duration of fixations as compared with Spanish 14. The older Spanishspeaking group surpassed the younger in all respects. Differences in individual performances at the end of Selection I are shown in Figures Ito 9. This selection was chosen because all the subjects made a satisfactory response when the blank was reached. The individual records of all the subjects used in the group comparisons are shown in these graphs. The subjects are arranged in rank order of reading age and the mean for the Spanishspeaking group is shown on each graph. It is possible, then, in these graphs to make both individual and group comparisons and at the same time to get a fair conception of the rank correlation existing between reading age and readiness of response at the end of a selection. Figures 1 and 2 show the number of fixations on the last line of Selection I for Spanish and English 14 and 11, respectively. Since the two graphs are on the same base and placed side by side it is possible to compare the records of any groups or individuals concerned with the records of any of the other groups or individuals. Figure 1 shows that the mean number of fixations for the last line of Selection I was less for Englishthan Spanidi-speaking subjects on the fourteen-year level, but it also shows that three of the best records were made by Spanish-speaking subjects as compared with one for the English-speaking subjects. There is no clear advantage in favor of the English-speaking groups so far as number of fixations is concerned. However, in Figure 2 it is clearly seen that the English-speaking subjects made a more favorable showing. When Figure 2 is compared with Figure 1, it is seen that the older Spanish-speaking group was superior to the younger. These facts seem significant in view of the findings for the number of fixations per line for material exclusive of the first and last lines which were shown for the different groups in Chapter 111 or which will be shown for individuals in a later discussion. For material exclusive of the first and last lines, the eleven-year-old Spanish-speaking group required fewer fixations per line than either of the English-speaking groups of the fourteen-year-old Spanish-speaking group. There may be some fine differences in degree of comprehension that are revealed by the analyses of eye-movement records that are lost when the subject makes written responses on a reading test, of there is the possibility that some phase of the language handicap exclusive of comprehension is responsible for these differences that seem to contradict other evidence that has been presented. An attempt at explanation in terms of the latter hypothesis will be presented in the final chapter. Figure 3 shows the individual records of the combined Spanish-speaking groups compared with the records of the accelerated English-speaking group with reference to the number of fixations for the last line of Selection I. Both groups are arranged in rank order of reading age, reading from left to right. It may be seen from this graph that only three of the twelve English-speaking subjects made more fixations than the mean for the Spanish-speaking subjects. A distinctly better showing was made by the English-speaking subjects. It will be recalled that the Spanish-speaking subjects made fewer regressive movements on the average than the English-speaking subjects, and that relatively large group differences were sometimes shown. The first and last lines were not counted. When comparisons were made for the last line of Selection I with reference to the number of regressive movements, it was found that the English-speaking groups made a better showing in every case. Also, it was found that Spanish 14 was superior to Spanish 11. These comparisons for individuals arranged in rank order of reading age are shown in Figures 4,5, and 6. Spanish-speaking subjects required significantly less time for each fixation on the material read before the camera where the record for the first and last lines were not counted. Wnen the records for the last line of Selection I were studied it was found that the Englishspeaking subjects made shorter fixations. Individual comparisons for all the groups are shown in Figures 7, 8, and 9. It is to be noted in these figures that the older Spanish-speaking subjects were more successful with reference to number of fixations and number of regressions. Figure 9 snows an interesting fact that may be significant in view of other findings. The upper 50 per cent of the English-speaking subjects, from the standpoint of reading age, made a better showing than the Spanidi-speaking subjects, but this standing was practically reversed for the lower 50 pe r cent. Relative Difficulties Experienced by Spanish- Speaking Subjects in Supplying a Word at the End of a Selection Written in English and a Parallel Selection Written in Spanish A summary of the differences in readiness of response shown by Spanish-speaking subjects in supplying a missing word at the end of a selection in English and a parallel selection in Spanish may be given since four of the subjects read two selections of the material in both languages Such a summary should furnish data that is useful in explaining the less favorable showing of the Spanish-speaking subjects where their records for leading the last line of Selection I were compared with the records of Englishspeaking subjects. The material read ended with a quotation followed by the explanatory clause, “said the . . . , ” or in Spanish, ”dijo el .. . H The number of fixations, the number of regressive movements, and the total time for the last two words and the blank for these selections in the two languages are shown in Table XXVI The data of Table while based on the records of only four subjects, indicate that Spanish-speaking subjects respond more readily in Spanish than in English. It is to be remembered in this connection that better eye-movement habits were shown by these subjects for reading English than for reading Spanish. It should be added that the significance of these particular measures was not foreseen during the experimental period and some objection might be raised to the controls that were used. The subjects read the material in English and a little over a month later read the same materia.l in Spanish. There is the possibility that some of them remembered what had been read previously. Effect of Change in Difficulty Of Reading Material on the Eye-Movement Habits of Spanish-Speaking Subjects When the eye-movement records for the various groups were presented in Chapter 111, it was shown that the number of fixations, the number of regressive movements, and the total time increased with increases in the difficulty of the reading material. This was true for all the groups. Individual subjects were affected in different ways. The same selection did not offer the same relative difficulty to all subjects. In addition, all <f the subjects did not comprehend every selection, which reduces the number of usable records very materially for some selections. Under these conditions, it is impossible to isolate differences in maturity of eye-movement habits which result solely from changes in the difficulty of material. Individual Performances for the Various Eye- Movement Measures Summaries of individual eye-movement records for Selection I are presented graphically in Figures 10 to 18. Little discussion will be required for these graphs after their general plan is explained. The subjects are all arranged in rank order of reading age beginning at the left of each figure. In each graph the records for a Spanish-speaking group and an English-speaking group is drawn on each graph. The fourteen- and eleven-yearold subjects are always presented on the same page and with the same base line. This makes comparisons between any combination of groups or individuals possible. The average number of fixations per line is shown in Figures 10 to 12. The Spanish—spea.king subjects made fewer fixations when considered as groups but considerable individual variation is shown. Spanish 11 showed a wider range than any other of the groups. Apparently there is a high correlation beteen number of fixations per line and reading age for this group. In most cases it is possible to tell from the graphs whether the means are really representative of group performances. Where they are not, the number of individuals above or below the mean may have little significance. A better procedure in such a case is to compare each individual of a group with the individual of the same rank in reading age in another group. For example, on the eleven-year level only one Spanish-speaking subject made more fixations per line than the English-speaking subject of the same rank. On the fourteen-year level there are three subjects above and three below the corresponding subjects of the same age and rank, but the English-speaking subjects were considerably higher in two cases. The Spanish-speaking subjects snowed shorter fixation time than either the retarded or the accelerated English-speaking groups. These records are summarized in Figures 13 to 15. It will be noticed that the mean for Spanish 14 seems higher than it should be. This can be explained by reference to Figure 10 where it will be seen that the subjects above the mean in average duration of fixations made more fixations per line. The number of Spanish-speaking subjects who made briefer pauses than the English-speaking subjects of the same reading age rank were: 4 out of six on the fourteen-year level, 5 out of six on the eleven-year level and 7 out of twelve when compared with the accelerated group. Spanish- 11,as a group, had a lower mean than Spanish 14, but individual comparisons fail to confirm their superiority with reference to this measure. The most apparent differences are shown with reference to the average number of regressive movements per line, in Figures 16 to 18. The Spanishspeaking subjects made fewer regressive movements than the English-speaking subjects. The curves for the upper levels show few intersection points but this is not true when the Spanish-speaking groups are compared with the accelerated English-speaking group. Summary of Results of Individual Comparisons The small number of eight-year-old subjects made it possible to report individual records for them when the various groups were compared with reference to their eyemovement habits. However, mere quantitative statements about an eye-movement record can not entirely take the place of an analysis of the record itself. Two records for eight-year-old subjects who were still in the ’’deciphering” stage were presented, one record for each language group. The record for the English-speaking child showed more “confusion periods”, but parts of it gave evidence of a more mature type of eye-movement habits than was shown for the Spanish-speaking child. Two records were presented for subjects with a higher order of eye-movement habits, and in this case the record of the English-speaking child was distinctly superior. When compared either as groups or individuals English-speaking subjects on the eight-year level showed more mature eyemovement habits than Spanish-speaking children who were of equal reading ability as measured by reading tests, and who were, on the average, superior in mental ability, as measured by a non-language mental test, and in arithmetical computation. Fewer regressive movements were made by fourteenyear-old Spanish-speaking children when they were paired subject for subject with eleven-year-old subjects of the same reading age. There seemed to be growth with age in respect to this measure. This tendency was not found for English-speaking subjects except in cases where marked immaturity of eye-movement habits was shown. The records of a Spaaish-speaking subject, aged 11, were compared with the records of five English-speaking subjects who had approximately the same reading age but were from several different age levels. Four eye-movement measures were considered for each of two selections. The records of the Spanish-speaking child compared favorably with those of the English-speaking children. Spanish-speaking subjects were found to read a selection where ”Age“ and ”Youth w were personified as well as English-speaking subjects on the fourteen-year- level and better than the eleven-year-old group or the accelerated English-speaking groups. The rank order of the groups for the unfamiliar word ’’planet*’ with reference to the eyemovement measures was: (1) English 14, (2) Spanish 14, (3) English 9.7, (4) Spanish 11, (5) English 11. The Spanish-speaking subjects failed to fill the blank at the end of a selection as readily as the Englishspeaking subjects. The older subjects made more favorable showing than the younger ones. Individual summaries with reference to the eye-movement measures showed that Spanish-speaking subjects tend to have better eye-movement habits than either retarded or accelerated English-speaking subjects of the same reading age. It is to be pointed out that since the Spanish-speaking subjects are retarded they are necessarily compared with retarded subjects when age is held constant. Eleven-year-old Spanish-speaking subjects have better eyemovement habits than fourteen-year-old subjects with respect to three of the four commonly used eye-movement measures. 21 ... , See Figures 10 and 11 below. Plate n. Silent Heading by Subject ECBB Plate VII. Silent Reading of Subject MFEB. Plate VIII. Silent Reading by Subject TGSB Plate IX. Silent Reading by Subject WBEB Plate X. Silent Reading by Subject JRSI4 of Selection TV Plate XI. Silent Reading by Subject CLSII of Selection IV. Plate XII. Silent Reading by Subject POSI4 of Selection 11. Plate XIII. Silent Reading by Subject CPSII of Selection 11. Palte XIV. Silent Reading by Subject SESI4 of Selection 111. Plate XV. Silent Reading by Subject TCSII of Selection 111. Plate XVI. Silent Rending by Subject ACSII of Selection 111. Flat® XVII. Silent Reading by Subject LJIII of Selection 111. Plate mil. Silent Reading by Subject E 558.3 of Selection 111. Plate XIX. Silent Reading by Subject CSSI4 of Selection 111. Plate XX. Silent Reading by Subject CAKE9.B of Selection 111. Plate XXI. Silent Reading by Subject JDEI4 of Selection 111. Plate XXII. Example of a Ready Response at the End of a Selection Plate XXlll.Example of Difficulty at the End of a Selection Plate XXIV.. Fairly Mature Performance at the End of a Selection Plate XV.. Less Mature Performance at the End of a Selection FIG. I NUMBER OF FIXATIONS FOR LAST LINE OF SELECTION I — SPANISH 14 AND ENGLISH 14 FIG. 2 NUMBER OF FIXATIONS FOR LAST LINE OF SELECTION I SPANISH II ANO ENGLISH If FIS. 3. NUMBER OF FIXATIONS FOR. LAST L/IME OF SELECTION I SPANISH II AND (4 AND FMSLISH 9.7 FIG. 4. NUMBER OF REGRESSIVE MOVEMENTS FOR LAST LINE OF SEL6CTION I —SPANISH H AND ENGLISH II FIG. 5. NUMBER OF REGRESSIVE MOVEMENTS FOR LAST UNE OF SELECTION I — SPANISH l( ANO ENGLISH II FIG.fo. NUMBER. OF REGRESSIVE MOVEMENTS FOR LAST LINE OF SELECTION I SPANISH It AND M- AND ENGLISH a/ FIG. 7. AVERAGE DURATION OF FIXATIONS LAST LINE OF SELECTION I — SPANISH 14 AND ENGLISH H FIG.AVERAGE DURATION OF FIXATIONS LAST LINE OF SELECTION I — SPANISH II ANO ENGLISH II FIG 9 AVERAGE PMRATION CF FIXATIONS — LAST LINE OF SELECTION Z SPANISH H AMP IA AND ENGLISH 9-7 FIG. 10 AVERAGE NUMBER OF FIXATIONS PER LINE SPANISH 14 AND ENGLISH 14- FIG. II . AVERAGE NUMOBR OF FIXATIONS PER. L INE SPANISH II AMP ENGLISH II FIG. IZ. AVERAGE NUMBER. OF FIXATIONS PER LINE SPANISH II AND 14 COMPARED WITH ENGLISH 9.7 fig, 14. average duration <»p fixation SPANISH H ANO ENGLISH II FIG- IS AVERASE- DURATION OF FIXATION SPANISH I* AKO ENGLISH »4 FIG-15 Average duration of FIXATION SPANISH If AND 1+ ANDENQMSH 3.7 FIG. 16. AVERAGE HUMBER OF REGRESSIVE MOVEMENTS PER LINE — SPANISH 14 ANP ENGLISH/4 FIG-17. AVERAGE NUMBER OF REGRESSIVE MOVeMEHTS FER. LINE —* SPANISH l« ANO ENGLISH II FIG. >6 . AVERAGE NUMBER OF REGRESSIVE MOVEMENTS FER. LINE SPANISH I) AND 14 AND ENGLISH 97 Measures Subjects TGS8 WBE8 Average time per line 83,85 67.75 Average number of fixations per line 7.75 7.00 Average duration of fixations 10.73 9.67 Average number of regressions per line 1.25 1.00 TABLE XX SUMMARY OF RECORDS OF SUBJECTS TGSB AND WBEB Subjects Read- ing Age Pic- ture supple- ment score Vo- cab- ulary score Com- pre- hen sion score Rate Score Care- less- ness score* Men- tal index Arith. compu- tation age JRS 14 146 5.6 93 46 132 57 —— 110.0 CLS 11 145.1 r.3 92 42 148 23 76 163.0 PCS 14 127.6 6.0 72 38 96 142 34 153.5 CPS 11 130 4.0 74 40 102 142 38 127.5 SHS 14 127 5.6 50 42 108 41 129.0 TCS 11 127 8.0 53 40 99 25 46 130.0 TABLE XXI TEST RECORDS OF SUBJECTS USED IN MAKING INDIVIDUAL COMPARISONS Subjects Average Time Per Line Average Number of Fixations Average Duration of Fixations Average Number Regressive Movement s Per Line JRS14 53.40 7.60 7.02 .60 CLS11 53.60 8.60 6.23 1.20 PCS14 59.33 9.33 6.35 .50 CPS11 54.00 8.16 6.61 1.00 SHS14 109.75 12.75 8.60 2.75 TCS11 57.00 10.50 5.42 2.00 TABLE XXII SUMMARIES OF EYE-MOVEMENT RECORDS FOR THREE PAIRS OF SPANISH-SPEAKING SUBJECTS AGES 11 AND 14 ACS11 LJE11 Subjects CANE9.8 JDE14 ESE8.3 CSE14 Reading Age 128.00 131.00 129.60 129.00 128.00 128.00 Selection L I Average time per line 63.25 64.50 102.50 87.5 43.25 81.00 Average number of fixations per line 8.75 9.50 14.50 13.00 6.50 9.50 Average duration of fixations 7.22 6.78 7.06 6.73 6.65 8.52 Average number of regressive movements 1.00 1.50 4.75 4.25 1.00 2.50 Selection II Average time per line 58.33 81.83 65.50 73.50 38.60 99.83 Average number of fixations per line 8.66 11.16 10.00 10.83 6.00 10.00 Average duration of fixations 6.75 7.32 6.55 6.78 6.40 9.98 Average number of regressive movement s 1.00 2.50 3.00 2.50 1.00 2.33 TABLE XXIII COMPARISON OF THE RECORDS OF AN ELEVEN-YEAR-OLD SPANISH-SPEAKING SUBJECT WITH THE RECORDS OF ENGLISH-SPEAKING SUBJECTS WHO HAVE THE SAME READING AGE Measures Groups Sp.14 Eng.14 Sp.ll Eng.11 2ng.9.7 "Age" Average duration 7.63 8 .66 7.45 8.23 7.22 Average number per subject 3.66 3.00 3.33 4.23 4.09 Average time per subject 28.00 26.00 24.83 35.66 29.54 "Youth" Average duration 5.85 8.41 6.17 8.00 7.52 Average number per subject 3.50 2.83 3.83 3.66 3.45 Average time per subject 20.50 18.00 23.66 29.33 26.00 "planet" Average duration 5.90 6.80 5.30 5.92 7.07 Average number per subject 1.83 .83 2.16 2.16 1.27 Average time per subject 10.83 5.66 11.50 12.83 9.00 TABLE XXIV COMPARISON OF THE DIFFERENT GROUPS OF SUBJECTS WITH REFERENCE TO SPECIAL WORD DIFFICULTIES Group Average Number of Fixations Average Number of Regressions Average Duration of Fixations Spanish 14 10.16 3.33 7.77 English 14 11.00 2.66 7.50 Spanish 11 10.33 2.66 8.40 English 11 8.66 2.00 8.69 English 9.7 9.00 2.16 8.14 TABLE XXV SUMMARY OF GROUP PERFORMANCES ON THE LAST LINE OF SELECTION I Spanish English Differ- ence Per Cent Excess of English Over Spanish Number of Fixations 22 23 1 4.50 Number of Regressive Movements 5 7 2 40.00 Total Time 171 249 68 45.61 Average Duration of Fixations 7. 77 10.82 3.05 39.25 TABLE XXVJt DIFFERENCE IN EYE-MOVEMENT RECORDS FOR SPANISHSPEAKING SUBJECTS FOR SUPPLYING A WORD WHICH HAS BEEN LEFT OUT AT THE END OF A SENTENCE CHAPTER V RELATIONSHIPS FOUND BETWEEN VARIOUS MEASURES OF THE READING HABITS OF SPANISH- SPEAKING SUBJECTS The data showing the results of a study of the reading habits of Spanish-speaking children have been presented for groups and for individuals. Different measures have been taken up separately and comparisons made with respect to them. It seems profitaole in this chapter to show something of the likelihood that subject who stands high with respect to one measure will tend to hold his position with respect to other measures. The number of subjects used in the study was necessarily restricted, for practical reasons, to an extent that, in many cases, precluded reliable measures of relationship. For this reason coefficients of correlation will only be shown when no other convenient method can be devised to show relative standings of the subjects on different measures used in the study. When correlation coefficients are shown they are to be taken as indications of tendencies and trends rather than precise measures. Relative Standings of the Spanish-Speaking Subjects on Eye-Movement Measures from One Selection to Another It is a well-known fact that changes in the difficulty of the material read affects the eye-movements of the reader. The material used before the eye-movement camera was progressively more difficult from selection to selection. Under these conditions it is expected that as a rule the subjects would make more fixations and more regressive movements, and require a longer time on any given selection than on the previous selection. However, if all the subjects preserved their normal reading habits throughout all the selections there should be little change in the relative standings of individual members of the groups. Figures 19 to 22 show relative standings for twelve Spanish-speaking subjects for Selections I, 11, and 111. The subjects are arranged in rank order of reading age beginning at the left. The figures show the correlation between performances on the three selections and also something of the relationship between reading age and the various eye-movement measures. Figure 19 shows the small changes in relative standing of the different subjects with reference to average time per line for the three selections. All of the curves have the same general shape. It will be noticed that the selections presented the same order of difficulty for most of the subjects. Rank correlatiors which agree very closely with the predictions that would be made from a study of the graph are given in Table XXVII Figure 20 shows relative standings of the subjects on the seme material as measured by the average number of fixations per line. The rank relationships are again seen to be very close. The general tendency may be seen in this graph for the number of fixations pr line to increase as reading age decreases, but the number of exceptions is great enough to reduce the correlation when it is expressed as a rank coefficient of correlation. Relative standings with reference to the average duration of fixations are shown in Figure 21. The curves in this graph are fairly similar, but little improvement is shown with increase in reading age. The standings for average number of regressive movements per line are shown in Figure 22. Again there is seen to be a distinct tendency for subjects to hold their same relative positions for the three selections. Intercorrelations for the various eye-movement measures for the first three selections are shown in Table XXVII. The records of 13 Spanish-speaking subjects were used. It may be seen from Table XXVlTthat the coefficients of rank correlation show the same tendencies that were shown graphically in Figures 19 to 22. The lowest correlation were found with reference to average duration of fixations. This is to be expected since a subject may react to a change in difficulty by retaining his normal recognition unit and increasing the duration of his fixations, or he may reduce his recognition unit and show no increase or even show a reduction in the average duration of fixations. Relationship Between Different Eye- Movement Measures in the Case of Spanish-Speaking Subjects A subject may make a favorable showing with respect to some eye-movement measures and a poor showing on others, despite the fact that the various measures are interrelated. Total time, for example, may be analyzed into the number of fixations and the duration of fixations. So correlations between total time and either of these measures are spurious. Regressive movements are fixations, so if it is desired to find the relationship between the number of fixations and the number of regressive movements, the calculation must be made in terms of the total number of fixations less the number of regressive fixations. Regressive movements tend to increase the average duration of fixations. This was found to be true for all the groups included in the study, but no reason can be seen why a correlation between these two measures should be spurious. The rank correlation between the number of regressive movements and the average duration of fixations was found to be P* .39 t .16; and between number of fixa-22 tions and average duration of fixations P- .35 ± .17. These coefficients are too low, as compared with the size of their probable errors, to have much significance. This lack of correlation may be seen by again referring to Figures 20, 21, and 22. It is readily perceived that the curves for average duration of fixations (Figure 21) do not have the same general shape as the other curves. The low correlation between average duration of fixations and the average number of fixations seems to agree with the findings of Buswell quoted above. Of the low correlation between these two measures, he says: ”The significance of this practically zero correlation is that the two elements are almost entirely independent, which means that a wide recognition span may be accompanied by a long, medium, or short fixation time. It is possible, therefore, for a reader to make normal progress in one element while the other may be greatly under or over-developed.” 23 The Relation Between Test Scores and Eye- Movement Habits It was shown in Chapter IV that a subject may have considerable skill in getting meaning from printed material, as evidence by creditable scores on reading tests, and still have immature eye-movement habits. Plate XIV illustrates this fact. On the other hand, even the younger readers may show fairly mature eye-movement habits on material within their range of difficulty. It is certain, therefore, that perfect correlation is not to be expected between test scores and eye-movement haOits, but some relationship does exist. The curves that were presented in Figures 19 to 22 show this in a rough way. Since a number of reading tests were given in selecting the subjects used in this study, it is possible to show rank correlations between the different test measures and the different eye-movement measures. These correlations are shown in Table XXVIH. The data of Table XXVIII show that the best single test measure found for predicting the eye-movement habits of Spanish-speaking subjects is a vocabulary test. This measure was one of the tests that entered into reading age, which would suggest that lack of word knowledge is one of the most serious handicaps the Spanish-speaking child has to the development of his reading habits. It is to be noted that the highest coefficient of correlation was between vocabulary and regressive movements. This would tend to indicate that word difficulty is one of the important causes of regressive movements for Spanish-speak!ng subjects. The lower correlation for English-speaking subjects, P* . 472 ±.151, seems in agreement with other data that have been presented. The eye-movement measure which showed the highest correlation with test measures was the average number of fixations per line. The highest relationship between this measure and the test measures was found with respect to reading age. It was pointed out above that in Figure 20 a close relationship was shown between these measures. Summary Three kinds of relationships between different measures of the reading habits of Spanish-speaking subjects were presented in this chapter. First, it was found that Spanish-speaking subjects tend to keep their same relative standings with reference to eye-movement measures from selection to selection of increasing difficulty. Second, intercorrelations between eye-movement measures were found to be low, but it was shown graphically that some relationship exists between them. Third, the test measures were compared with the eye-movement measures. When this comparison was made it was found that a vocabulary test was the best single test measure for indicating eyemovement habits and that average duration of fixations showed the highest mean correlation with the test measures. 23 Buswell found the Pearson correlation between the average number of fixations per line and average duration of fixations for 83 subjects in Grades II to VI to be r< .49 * .056, and high school juniors -.08 ± .055. Buswell, G. T.: ’’Fundamental Reading Habits: A Suudy of Their Development,” op. cit. p. 107. 23 Ibid., p. 107. 24 Throughout the study subjects have been ranked in terms of desirability of position. So the expression of the rank correlations between eye-movement measures and test scores as positive coefficients means that those who showed desirable performances with respect to one kind of measures tended to do so with respect to the other kind also. If the size of the scores had been taken into account such coefficients would be negative. SUBJECT'S FIG. 19 RELATIVE STAMPING WITH RESPECT TO AVERAGE TIME PER LINE FOR THREE SELECTIONS PF INCREASING DIFFICULTY — TWELVE SPANISH SPEAKING SUBJECTS ARRANGED IK RANK. ORDER. 0F REAPING AGE FIG ZO AVERAGE NUMBER OF FIXATIONS FER LINE FOR. THREE SELECTIONS OF INCREASING DIFFICULTY — TWELVE SPANISH SPEAKING SUBJECTS FIG £1 AVERAGE DURATION OF FIXATIONS FOR THR.EE SELECTIONS OF INCREASING DIFFICULTY TWELVE SPANISH SPEAKING SUBJECTS F1G2.2 AVERAGE NUMBER. OF REGRESSIVE MOVEMENTS PER LINE PORTHREE SELECTIONS OF INCREASING DIFFICULTY — TWELVE SPANISH SPEAKINO SUBJECTS Average Time per Selection II Line Selection III Average Time per Line I II .914 ± .032 .873 ± .046 .855 t .052 Average Number Selection II of Fixations Per Line Selection III Average Number of Fixations per Line I II .915 ± .032 .869 ± .047 .838 t .058 Average Duration Selection II of Fixations Selection III Average Durat ion of Fixations I II .446 i .156 .7414 .088 .675 t .106 Average Number Per Line Selection II of Regressive Movements Selection III Average Number of Regressive Movements per Line I II .058 .837 i .058 .502 t .146 TABLE XXVII RANK DIFFERENCE INTERCORRELATIONS BETWEEN EYE- MOVEMENT MEASURES FOR THREE SELECTIONS OF INCREASING DIFFICULTY (13 Spanish-Speaking Subjects) Eye-Move- ment Measures Test Measures Reading Age Picture Supplement Vocabu- Compre- hension Mean lary Speed Average Time per .584 .393 .555 .480 .307 .463 Line Average Num- ber of Fixa- ±.129 ±.165 ±.135 ±.150 ±.177 tions per .638 .496 .540 .532 .347 .530 line Average ±.116 ±.146 ±.136 ±.137 ±.168 Duration of .343 • 248 .588 .183 .400 .352 Fixations 1.172 ±.183 ±.128 ±.189 ±.205 Average Number of .473 .455 .647 .321 .205 .420 Regressive Movement s per Line ±.152 ±.156 1.113 ±.175 ±.186 Mean Coefficient .509 .398 .582 .379 .314 TABLE XXVIII RELATIONSHIP BETWEEN TEST MEASURES AND EYE- MOVEMENT MEASURES CHAPTER VI INTERPRETATION AND CONCLUSIONS The eye-movement technique was used in making a study of the reading habits of bilingual children. Eyemovement records were made for twenty-one Spanish-speaking subjects selected from the schools of Austin — nine at age 14, six at age 11, and five at age 8. E a ch subject on the upper age levels read four selections of increasing difficulty before the eye-movement camera. The eight-year-old group read one selection. Their records were summarized for group and individual comparisons and in some cases given individual analysis. Control groups of English-speaking subjects were selected of the same age and of the same reading ability as measured by standardized reading tests. Groups of six were paired at ages 11 and 14, and groups of five at age 8. A fourth control group of English-speaking subjects was selected with the same mean reading age as the combined eleven- and fourteen-year-old Spanish-speaking groups, but who averaged 9.7 years in chronological age. The experimental results were presented in Chapters 111, IV, and V. From these results it is possible to state some conclusions with reference to the eye-movement habits of the Spanish-speaking subjects used in this investigation. Conclusions Based on the Preliminary Study and Tests of Comprehension Before the Eye-Movement Camera (1) Spanish-speaking boys could not be found in the Austin schools who have made progress in reading equal to the progress to be expected of normal English-speaking children. Only one Spanish-speddng subject was found with normal age-grade placement. School records showed that Spanish-speaking girls were more seriously retarded, on the average, than Spanish-speaking boys, so it was not thought possible to get a more representative sampling through the use of both sexes. The results of this study agree with the results of other studies which show serious retardation for Spanish-speaking children in the public schools. (2) Spanish-speaking subjects received, credit for comprehension on approximately the same number of selections as English-speaking subjects of the same age and reading age. The most difficult selection used before the eye- movement camera was comprehended by a larger per cent of Spanish-speaking subjects than by an accelerated Englishspeaking group of the same mean reading age. When compared with English-speaking subjects of the same age and reading age, or of the same reading age, Spanish-speaking subjects make as favorable or more favorable showing with respect to comprehension of reading material. Conclusions Based on an Experimental Study of Groups (1) The results of this study indicate that Spanishspeaking children do not show growth in eye-movement habits between the ages of 11 and 14, except in the matter of reducing the average number of regressive movements per unit of material read. Marked growth with respect to all the eye-movement measures is found between the ages of 8 and 11. (2) Spanish-speaking subjects of the eleven- and fourteen-year levels require slightly less time, on the average, for reading a unit of material than Englishspeaking subjects of the same reading age and same age. The older group tends to fall below an accelerated English-speaking group of the same reading age and the younger group tends to require a little less time than the accelerated English-speaking group. Spanish-speaking subjects make fewer pauses per line than eleven-year-old English-speaking subjects who are retarded; approximately the same number of pauses as fourteen-year-old Englishspeaking subjects who are retarded; and more pauses per line, on the average, than accelerated English-speaking subjects with the same reading age. With respect to average duration of fixations and the number of regressive movements per line, Spanish-speaking subjects made records that are clearly superior to the records of either retarded or accelerated English-speaking groups with the same mean reading age. (3) On the eight-year level, the English-speaking subjects showed better eye-movement habits than the Spanish-speaking subjects in all respects except average duration of fixations. With respect to this measure, the chances were found to be 85 in 100 that Spanish-speaking subjects would make briefer fixation pauses. (4) Spanish-speaking subjects made a larger per cent of their regressive eye-movements of rhe type that results from failure to make a sufficiently long return eye-sweep from the end of one line to the beginning of the next line. (5) Regressive movements served to increase the average duration of fixations for all subjects. The per cent of increase was greater for English-speaking subjects than for Spanish-speaking subjects. (6) Spanish-speaking subjects tend to show less variaMlity in reading performance than English-speaking subjects, but the evidence is not conclusive, except possibly on the eight-year level. Coefficients of variability were calculated for individuals and for groups. Spanish-speaking subjects were found to be less variable on eleven out of eighteen measures. (7) Spanish-speaking subjects who can read material in both English and Spanish show more mature eye-movement habits while reading English, except that the duration of fixations is less while reading material in Spanish. (8) Bilingual readers show more variability with respect to average duration of fixations for material in Spanish than for material in English. Conclusions Based on Individual Analyses (1) Test records for eight-year-old Spanish-speaking subjects indicated a greater maturity in reading habits than would be predicted from a study of their eye-movement records. One eight-year-old Spanish-speaking child was found, however, who showed fairly mature eye-movement habits for material within his range of difficulty. (2) Eight-year-old Spanish-speaking subjects show less maturity in eye-movement habits when compared individual to individual with English-speaking subjects of the same chronological age ahd reading age. (3) Spanish-speaking subjects eleven years of age showed about the same general maturity in eye-movement habits as fourteen-year-old Spanish-speaking subjects when they were compared individual to individual with the same reading age. The older subjects, however, made a more favorable showing with respect to the number of regressive movements. (4) An eleven-year-old Spanish-speaking subject was found with approximately the same reading age as five English-speaking subjects, aged 8,9, 11, 14, and 14 respectively. The eye-movement records of the Spanish-speaking subject showed him to be inferior in general maturity of reading nabits to the nine-year-old Englishspeaking subject but superior to the other members of the group with whom he was compared. (5) The eye-movement records indicate that Spanish-speaking subjects read a selection containing personi fications as well as English-speaking subjects. (6) Spanish-speaking subjects show less readiness in the recognition of difficult words than English-speaking subjects, despite the fact that they made slightly higher scores on vocabulary tests. (7) Eye-movement records indicate less maturity in reading habits for bilingual subjects than unilingual subjects while supplying a word to complete the meaning of the last sentence of a selection Older subjects are more successful in this respect than younger subjects. (8) Bilingual readers supply a word at the end of a selection in Spanish more readily than when the same material is in English. (9) Spanish-speaking subjects show no changes in reading habits with changes in the difficulty of the reading material that would distinguish them from any other group of readers. Conclusions Based on a Comparison of the Various Measures Used in the Study (1) In reading selections of increasing difficulty there was a decided tendency for the subjects of this study to maintain the same relative standings from selection to selection with respect to all the eye-movement measures. (2) Correlations between eye-movement measures for Spanish-spesking subjects are approximately the same as correlations found for other subjects. (3) The results of this study indicate that the test measure having the closest agreement with eyemovement measures for Spanish-speaking subjects is a vocabulary test. Interpretations Some of the conclusions presented above need some explanation and interpretation. In some cases objective data were collected which seem to be more or less contradictory in their implications. The purpose of this section is to clear up such cases where it is possible. The first conclusion based on the study of group performances was to the effect that growth is shown for only one eye—movement measure, the number of regressive movements, for Spanish-speaking subjects from age eleven to age fourteen. But evidence of growth was revealed by the technique which does not appear in a simple statement of total time, number and duration of fixations, or the number of regressive movements. This evidence may be summarized by saying that the older subjects showed more maturity in reading difficult words and in supplying a word at the end of a paragraph. So there is evidence of growth from eleven to fourteen, and it is entirely possible that growth in reading habits would have been found if twelve- or thirteen-yearold subjects had been used. Eliminations from school as a result of economic pressure are more likely to occur on We higher age level, and children who have made the best school adjustment would have the best chance to secure employment. The general tendency that was exhibited for Spanishspeaking subjects to show more mature eye-movement habits than English-speaking subjects may also required a word of explanation. It was pointed out in Chapter I that the Spanish-speaking subjects made a more favorable showing with respect to both intelligence and arithmetic tests than the English-speaking subjects of the same age. The Spanish-speaking subjects may have developed eye-movement habits above their reading ages. The accelerated group of English-speaking children did not test below the Spanish-speaking groups in any respect, but there may be some relationships between physiological growth and growth in eye—movement habits that have not been investigated as yet. Spanish-speaking subjects showed fewer regressive movements per unit of material read than English-speaking subjects. It is debatable whether regressive movements for subjects who have not approached the adult level of maturity in reading are signs of immaturity or sign of a period of transition to a higher order of reading habits. Since it was shown that regressive movements increase the average duration of fixations, the fact that Englishspeaking subjects made more regressive movements may serve partially to explain their unfavorable showing with respect to average duration of fixations. The fact that Spanish-speaking subjects showed less variability than English-speaking subjects with reference to some of the eye-movement measures has some implications of the language handicap. The reader who has ease and fluency in a language is able in many cases to fixate only the key words of a line of material. He is able to supply some words or acceptable synonyms for them without seeing them directly. As a result he might show considerable variability in his performance. The reader who is unfamiliar with a language would have to see each word in a direct manner which would tend to break each line into approximately the same number of recognition units. Since he sees each word he is less likely to make regressive movements and as a result, he would show less variability. The less favorable showing of Spanish-speaking subjects on rhe last line of Selection I, or where they were required to supply a word to complete the meaning of the last sentence of the selection, has at least two possible explanations. The first of these explanations is based on the possibility that there may be fine differences in the degree of comprehension a subject may show when he reads, and that his eye-movements are symtomatic of the mental processes involved. This explanation would seem to apply to individual differences between subjects regardless of language difficulties. If it were accepted as explaining the uniformly better showing of the Englishspeaking groups in filling the blank at the end of a selection, it would have to be said that despite a seeming superiority in eye-movement habits, the Spanishspeaking subjects miss some of the meaning in the material they have read and are consequently less successful in making a response which tests their comprehension. Another possible explanation of the differences between the two language groups can be found in a language handicap itself. The subject may have a fairly large recognition vocabulary, but when he is called upon to make a response in English, the response may be blocked by a more familiar Spanish word if he is accustomed to thinking in Spanish rather than in English. This explanation could be made more general by simply saying that for all subjects recognition is easier than reproduction or recall. Then it could be said that for the bilingual groups the differences between recognition time and overt verbal reaction greater than they were for the unilingual groups. The fact that bilingual readers supplied a word at the end of a selection in Spanish more readily than for a parallel selection in English tends to support the hypothesis that the ]e ss favorable showing of these subjects in making a response at the end of a selection of reading material when compared with unilingual subj ects with no better or even poorer eye-movement habits may be due to the fact that the English word is blocked by a Spanish word. But since one of the hypotheses offered does not exclude the other, it may be that these subjects are handicapped with reference to both precision of recognition and the readiness of making responses. Final Conclusions On the basis of the data which have been presented in this study the following general conclusions are made with reference to the reading habits of the subjects used: (1) Eight-year-old Spanish-speaking children show less maturity in eye-movement habits than should ordinarily be expected for children who have no language handicap. (2) Eleven- and fourteen-year-old Spanish-speaking children have on the average formed eye-movement habits which show a considerable degree of reading maturity. Some few individuals approach adult standards with reference to some of the measures. (3) Fourteen-year-old Spanish-speaking subjects show evidences of greater maturity in reading habits than eleven-year-old subjects of the same language group in three respects: the average number of regressive movements per line, readiness in the recognition of difficult words, and supplying a word to complete the meaning of a sentence. (4) Eight-year-old Spanish-speaking subjects made a less favorable showing than English-speaking subjects of the same age and reading age on all the eye-movement measures except average duration of fixations. (5) When eleven- and fourteen-year-old Spanishspeaking subjects are compared with English-speaking subjects of their own age and reading age or with accelerated subjects with the same reading age, they show more mature eye-movement habits. They show less maturity in eye-movement habits with respect to the recognition of difficult words and in the matter of supplying a word to complete the meaning of the last sentence in a selection of reading material. (6) The Spanish-speaking child does not respond in the same manner as the English-speaking child to the more subtle phases of the English language. BIBLIOGRAPHY Buswell, G. T.: A Laboratory Study of the Reading of Modern Foreign Languages, Vol/2, Publications oF the American an 3. Canadian Committees on Modern Languages. Macmillan, 1927, xii and 100 pp. --- - - : Fundamental Reading Habits: A Study of Their Development, 3uopTementarv~ Education Mo no gr aphs, No. 21,1922, 150 pp. Garretson, 0. K.: ”A Study of Causes of Retardation Among Mexican Children in a Small Public School System in Arizona,” Journal of Educational Psychology. 19:31 - 46. ~ January 1523. Manuel, H. T.: The Education of Mexican and Spanish- Speaking Children in Texas, The Fund for Research in the Social Sciences, The University of Texas, 1930, ix and 173 pp. Pavlov it ch, M.: Le language enf ant in. Acquisition de Serbe et de Francais par un enfant serbe. Paris’: Champion, l§2d. Pp. 2d4. Roquet, J.: Le develoypement du language observe chez un enTant bi 11ngue. Faris: dhambion, 1923 Saer, D.J. : ’’The Inquiry into the Effect of Bilingualism upon the Intelligence of Young Children,” Journal of Experimental Pedagogy. 1922, 6:233-240 and 553-^74. Smith, F.: ’’Bilingualism and Mental Development,” British Journal of Psychology, 1923, 12:270-282. Tireman, L. S.: ’’Reading in the Schools of New Mexico,” Elementary School Journal, 1930, 30:621 - 626. Yoshioka, J. G.: ”A Study of Bilingualism,” Pedagogical Seminary & Journal of Genetic Psychology, 1925. ooT ~ ' — APPENDIX A ♦ Reading age and arithmetical computation age are expressed in months; Picture supplement scores are the average of scores on three forms of the Burgess Picture Supplement Test; speed scores represent the total number of items attempted on two forms of the New Stanford Reading Test, Paragraph Meaning;Comprehension scores are Thorndike-McCall T-Scores; and vocabulary scores are total number right on two forms of the New Stanford Reading Test, Word meaning. Sub- ject Age Group Reading Age Pic- ture Supple- Com- pre- hen- sion Vocab- ulary Arith. Computa- tion Age. me nt Speed CH 14 149.6 8.3 138 49 85 146 JR 14 145.6 5.6 132 46 93 110 PC 14 127.6 6.0 96 38 72 153.5 SH 14 127.0 5.6 108 42 50 129.0 FR 14 121.6 4.3 99 37 52 136.0 FE 14 120.0 5.0 94 36.5 48 BC 14 119.0 3.6 68 39 83 129.0 CD 14 111.0 3.0 ■ » — 130.5 BG 14 109.0 2.0 69 36 86 MM MB CL 11 145 10.3 148 42 92 163 CP 11 129 4.0 102 40 74 127.5 AC 11 128 6.0 116 39 70 134.0 TC 11 127 8.0 99 40 53 130.0 OS 11 110 2.6 82 32 44 105.0 LC 11 103 2.3 86 32.5 32 112.5 EC 8 103.2 107.5 FG 8 91.8 99.0 FR 8 92.0 100.5 TG 8 88.0 •mb mm MM M» MM MM MM MM AA 8 87.2 91.5 LIST OF SUBJECTS AND TEST SCORES SPANISH-SPEAKING SUBJECTS Pic- Oom- Arith. ture pre- Vocab- Computa- Sub- Age Reading Supple- hen- ulary tion ject Group Age ment Speed s ion Age. RR 14 152.6 8.0 146 51 91 151.0 BM 14 152.3 9.3 139 51 92 139.5 JG 14 136.0 7.3 130 43 58 150.5 OS 14 124.3 6.6 101 42 48 137.0 JD 14 128.6 4.0 68 45 52 150.0 DC 14 117.0 3.0 106 39 40 117.0 EJS 11 139.0 6.6 108 46 74 136.0 AM 11 135.6 7.6 120 45 69 131.5 LJ 11 131.0 6.6 102 43 50 —» —• •• ML 11 125.0 5.3 84 38 63 135.0 RLG 11 122.3 4.3 95 36 54 126.0 RS 11 122.0 3.0 74 38 39 124.5 AN 11 105.0 1.3 — M 34 — 103.0 TW 9.7 154.6 9.0 127 54 89 126.5 AF 9.7 139.6 7.0 106 46 73 129.5 PP 9.7 137.0 9.0 137 48 71 127.5 PB 9.7 132.0 9.0 127 42 74 126.5 LLB 9.7 132.0 9.0 93 43 70 126.0 DE 9.7 131.6 6.0 91 42 63 128.5 ES 9.7 129.6 5.0 146 43 59 126.5 CAN 9.7 128.0 8.0 92 38 59 128.5 LC 9.7 122.3 3.6 80 38 44 131.5 ws 9.7 120.3 6.0 85 39 43 116.5 LJ 9.7 118.6 2.0 65 39 38 116.0 JK 9.7 109.0 2.0 76 32 38 115.0 MF 8 97.4 96.0 MC 8 95.9 88.0 RB 8 94.3 83.0 WB 8 87.8 88.0 JDB 8 86.4 88.0 ENGLISH - SPEAKING SUBJECTS APPENDIX B ♦Since only one selection was used for the 8-year old groups, no record could be counted unless satisfactory comprehension was shown* The lists of subjects do not have to be repeated here* Spanish- Speaking Subjects Selections English Speaking Subj ects Selections CH 1,2,3,4 RR 1,2,3,4 JR 1,2,3,4 BM 1,2,3,4 PC 1,2,3,4 JG 1,2,3,4 SH 1,2,3,- CS 1,2,3,- FR 1,2,3, — JD 1,2,3,4 FE 1,2,3,4 DC 1,2,3, — BC 1,- 3,4 CD 1,2,3, BG 1,2,- - CL 1,2,3,4 EJS 1,2,3,- CP 1,2,- - AM 1,2,3,4 AC 1,2,3,4 LJ 1,2,3,— TC 1,2,3,4 ML 1,2,3,- CS 1,2,3,4 RLG 1,2,3,- LC 1,- 3,- RS 1,-3, 4 AN 1,- 3,- TW 1,2,3,— AF 1,2,3,- PP 1,2,3,4 PB 1,2,3,- LLB 1,2,3, — DE 1,2,3,- ES 1,2,3,- CAN 1,2,3,4 LC 1,2, 3,4 ws 1,- 3,- LJ 1,- 3,- JK 1, LIST OF SUBJECTS AND SELECTIONS ON WHICH A SATISFACTORY DEGREE OF COMPREHENSION WAS SHOWN ♦ APPENDIX C Sub- ject Age Group I Selection II III IV GIS Adult 30.00 35.83 50.25 39.4 CH 14 76.00 79.00 76.00 77.40 JR 14 38.75 57.33 47.50 53.40 PC 14 61.75 61.75 59.33 80.50 SH 14 83.75 86.18 109.75 114.20 FR 14 74.75 69.83 78.50 92.00 FE 14 77.50 89.33 94.00 85.00 BC 14 78.00 79.00 82,75 81.20 CD 14 157.25 162.33 182.75 157.40 BG 14 66.25 60.50 60.33 53.40 CL 11 40.00 44.50 52.25 53.60 CP 11 55.60 54.00 59.50 75.20 AC 11 63.25 58.33 67.50 64.80 TC 11 57.00 65.50 61.25 73.80 CS 11 71.50 76.00 66.50 LC 11 96.50 93.16 94.00 102.60 Special Selection EC 8 92.75 FG 8 129.00 FR 8 59.00 TG 8 31.00 AA 8 146.00 LIST OF SUBJECTS AND EYE-MOVEMENT RECORDS AVERAGE LINE PER LINE SPANISH-SPEAKING SUBJECTS Sub- ject Age Group Selection I II III IV L Adult 23.50 25.33 28.00 25.80 W Adult 36.00 23.50 26.75 29.60 RR 14 60.50 52.83 71.75 57.20 BM 14 53.00 83.52 67.25 89.80 JG 14 91.50 69.66 80,75 112.20 OS 14 87.50 73.50 76.50 73.00 JD 14 81.00 99.83 85.75 86.80 DC 14 94.00 76.33 65.25 87.60 EJS 11 67.25 71.66 76.00 103.00 AM 11 69.25 100.50 111.25 122.00 LJ 11 64.50 81.83 98.50 82.00 ML 11 79.25 75.16 104.50 134.40 RLG 11 87.75 80.66 108.75 111.00 R8 11 107.50 171.80 135.50 123.20 AN 11 223.25 — — — — — — — — •— TW 9.7 44.25 53.33 48.50 59.00 AF 9.7 54.50 97.50 82.00 111.40 PP 9.7 44.75 60.16 70.25 78.80 PB 9.7 46.25 51.33 59.75 64.80 LLB 9.7 42.75 50.66 66.25 63.40 DE 9.7 104.75 154.20 122.25 200.00 ES 9.7 102.50 65.50 103.75 141.00 CAN 9.7 43.25 38.60 55.25 57.80 LC 9.7 59.25 73.16 75.25 40.20 WS 9.7 96.00 149.99 100.00 109.00 LJ 9.7 57.75 75.00 97.50 75.20 JK 9.7 87.75 115.66 W MB M MB Special Selection MF 8 108.25 MC 8 162.50 RB 8 77.50 WB 8 67.75 JDB 8: 123.50 AVERAGE TIME PER LINE ENGLISH-SPEAKING SUBJECTS Sub- ject Age Group I Selection IV II III GIS Adult 5.25 6.50 8.75 6.20 CH 14 11.50 12.83 10.50 12.00 JR 14 7.50 6.50 6.75 7.60 PC 14 9.50 9.33 11.25 13.60 SH 14 11.75 11.00 12.75 14.20 FR 14 11.00 10.16 9.75 11.00 FE 14 11.50 13.00 13.25 13.60 BC 14 12.25 13.00 12.00 13.00 CD 14 17.50 20 .16 18.00 21.00 BG 14 9.00 8.33 8.00 8.20 CL 11 6.75 7.33 8.75 8.60 CP 11 9.00 8.16 9.25 11.60 AC 11 8.75 8.66 7.25 8.40 TO 11 10.50 11.66 10.25 12.40 CS 11 12.25 11.00 9.50 MB MB MB LC 11 15.00 13.66 12.75 14.00 Special Selection EC 8 13.50 FG 8 13.25 FR 8 14.75 TG 8 7.75 AA 8 9.12 AVERAGE NUMBER OF FIXATIONS PER LINE SPANISH-SPEAKING SUBJECTS Sub- ject ?Age Group I Selection II III IV L Adult 5.85 5.33 5.75 5.60 W Adult 6.00 4.16 4.25 4.60 RR 14 9.25 9.16 11.25 8.80 BM 14 8.00 8.50 8.50 13.20 JG 14 14.35 11.83 12.00 14.80 OS 14 13.00 10.83 9.25 10.40 JD 14 9.50 10.00 9.25 9.60 DO 14 11.35 10.33 7.50 11.40 EJS 11 9.00 9.83 7.50 12.60 AM 11 10.75 14.16 14.75 15.20 LJ 11 9.75 11.16 11.25 10.20 ML 11 13.00 12.83 17.50 17.20 RLG 11 12.35 11.66 14.00 14.16 RS 11 13.75 20.00 14.25 15.00 AN 11 11.59 — »■ •MB «MB TW 9.7 7.50 8.16 6.75 8.20 AF 9.7 8.50 14.33 12.25 15.20 PP 9.7 7.50 7.66 10.00 10.00 PB 9.7 8.35 8.83 9.25 10.20 LLB 9.7 7.25 7.83 9.75 10.40 DE 9.7 12.50 18.20 13.00 19.60 ES 9.7 14.50 10.00 13.00 16.00 CAN 9.7 6.50 6.00 8.25 9.20 LC 9.7 8.50 10.00 8.75 6.40 ws 9.7 13.75 16.99 11.00 12.20 LJ 9.7 12.75 11.00 14.25 11.00 JK 9.7 11.50 15.16 ■■ Special Selection MF 8 11.50 MC 8 15.25 RB 8 10.50 WB 8 7.00 JDB 8 12.00 AVERAGE NUMBER OF FIXATIONS PER LINE ENGLISH-SPEAKING SUBJECTS Sub- ject Age Group Selection I II III IV GIS Adult 5.71 5.51 5.77 6.35 CH 14 6.60 6.15 7,23 6.45 JR 14 5.16 7.64 7.03 7.02 PC 14 6.33 6.35 7.15 5.47 SH 14 7.14 7.83 8.60 9.02 FR 14 6.79 6.86 8.05 8.36 FE 14 6.73 6.87 7.09 6.25 BC 14 6.36 6.07 6.85 6.24 CD 14 9.14 8.04 10.10 7.44 BG 14 7.36 7.22 7.56 6.51 CL 11 5.92 6.06 5.97 6.23 CP 11 6.19 6.61 6.43 6.48 AC 11 7.22 6.75 8.37 7.71 TC 11 5.42 5.61 5.97 5.95 CS 11 6.97 6.72 7.00 LC EC FG FR TG AA 11 8 8 8 8 8 6.43 £ 6.78 7.49 Special Selection 6.87 9.73 9.44 10.73 9.12 7.32 AVERAGE DURATION OF FIXATIONS SPANISH-SPEAKING SUBJECTS Sub- ject Group Selection I II III IV L Adult 4.47 4.75 4.86 4.60 W Adult 6.00 4.16 4.25 4.60 RR 14 6.54 5.76 6.37 6.50 BM 14 6.42 8.35 7.91 6.80 JG 14 6.42 5.87 6.73 7.58 OS 14 6.73 6.78 8.27 7.01 JD 14 8.52 9.98 9.26 9.04 DO 14 8.35 6.85 8.70 7.68 EJS 11 7.47 7.28 9.80 8.17 AM 11 6.44 7.04 7.54 8.02 LJ 11 6.78 7.32 8.75 8.01 ML 11 6.09 5.98 6.05 7.81 RLG 11 7.16 6.91 7.76 7.60 RS 11 7.98 8.59 9.50 8.21 AN 11 19.25 — — TW 9.7 5.90 6.53 7.18 7.19 AF 9.7 6.41 6.10 6.69 7.32 PP 9.7 6.17 7.84 7.02 7.88 PB 9.7 5.60 6.53 6.45 6.35 LLB 9.7 5.89 6.46 6.79 6.09 DE 9.7 8.38 8.47 9.40 10.20 ES 9.7 7.06 6.55 7.98 8.81 CAN 9.7 6.65 6.40 6.69 6.28 LC 9.7 6.85 7.31 8.60 6.28 ws 9.7 7.16 8.90 9.09 8.93 LJ 9.7 6.68 6.81 6.81 6.65 JK 9.7 7.63 7.62 mb Mt Special Selection MF 8 9.41 MC 8 10.65 RB 8 7.16 WB 8 9.67 JDB 8 10.29 AVERAGE DURATION OF FIXATIONS ENGLISH-SPEAKING SUBJECTS Sub- ject Age Group Selection I II III IV GIS Adult 0 .66 1.75 .40 CH 14 2.50 4.00 1.25 3.00 JR 14 .75 .16 .50 .60 PC 14 .50 .50 1.50 2.60 SH 14 3.50 2.16 2.75 3.04 FR 14 1.00 1.33 .75 1.20 FE 14 1.75 2.83 1.50 1.80 BC 14 1.25 2.16 1.75 2.40 CD 14 4.50 6.16 5.00 4.60 BG 14 1.50 .83 1.50 1.00 CL 11 .50 .83 1.50 1.20 CP 11 1.75 1.00 1.50 2.20 AC 11 1.00 1.00 .50 1.00 TC 11 2.00 2.50 2.00 2.20 CS 11 2.75 2.83 2.00 LC 11 4.00 3.66 2.75 2.20 Special Selection EC 8 4.25 FG 8 3.50 FR 8 2.75 TG 8 1.25 AA 8 4.75 AVERAGE NUMBER OF REGRESSIVE MOVEMENTS PER LINE SPANISH-SPEAKING SUBJECTS Sub- ject Age Group Selection I II III IV L Adult .75 .83 .75 1.00 W Adult 1.50 .16 .75 .60 RR 14 2.25 3.33 3.25 2.20 BM 14 1.75 1.46 1.00 2.00 JG 14 2.25 2.00 2.25 3.00 OS 14 4.25 2.50 1.50 1.60 JD 14 2.50 2.33 2.25 2.40 DO 14 1.50 1.66 .50 1.80 EJS 11 1.25 2.00 .50 3.20 AM 11 2.25 3.33 2.50 2.80 LJ 11 1.50 2.50 1.00 1.60 ML 11 4.25 4.16 6.25 5.60 RLG 11 2.75 2.50 2.75 2.60 RS 11 3.25 5.20 3.75 3.40 AN 11 6.25 — — ■M* «■* TW 9.7 1.75 1.33 1.25 1.20 AF 9.7 1.75 4.83 2.50 4.80 PP 9.7 .25 .33 1.75 1.40 PB 9.7 2.25 1.66 2.85 2.60 LIB 9.7 .25 .83 1.00 1.40 DE 9.7 2.75 4.00 2.00 6.00 ES 9.7 4.75 3.00 3.75 5.20 CAN 9.7 1.00 1.00 1.00 2.00 LC 9.7 2.25 2.33 1.75 .60 ws 9.7 4.00 3.50 1.75 2.60 LJ 9.7 3.75 2.83 4.25 2.80 JK 9.7 3.75 3.66 «a» Special Selection MF 8 3.50 MC 8 4.50 RB 8 2.50 WB 8 1.00 JDB 8 3.75 AVERAGE NUMBER OF REGRESSIVE MOVEMENTS PER LINE SPANISH-SPEAKING SUBJECTS