ill li''i i|iisii,!iit-!ifl!iSI i LB CORNELL UNIVERSITY LIBRARY GIFT OF The Estate of Dr. Elsie Murray The original of tliis book is in tine Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924032705539 Cornell Universlly Library LB1051.F85 E9 olin RIVERSIDE TEXTBOOKS IN EDUCATION EDITED BY ELLWOOD P. CUBBERLEY PROFESSOR OF EDUCATION LELAND STANFORD JUNIOR UNIVERSITY RIVERSIDE TEXTBOOKS IN EDUCATION RURAL LIFE AND EDUCATION. By E. P. CuBBERLEY, Professor of Education, Leiand Stanford Junior University. $1.50 net. Postpaid. THE HYGIENE OF THE SCHOOL CHILD. By L. M. Terman, Associate Professor of Education, Leiand Stanford Junior University. Si -65 net. Postpaid. EVOLUTION OF THE EDUCATIONAL IDEAL. By Mabel I. Emerson, First Assistant in Charge of the George Bancroft School, Boston. $1.00 net. Postpaid. HEALTH WORK IN THE SCHOOL. By E. B. HoAG, M.D., Medical Director, Long Beach City Schools, Cal., and L. M. Terman. ^1.60 net. Postpaid. DISCIPLINE AS A SCHOOL PROBLEIVI. By A. C. Ferry, Jr., District Superintendent of Schools, New York City. $1.25 net. Postpaid. HOW TO TEACH THE FUNDAMENTAL SUBJECTS. By C. N. Kendall, Commissioner of Education for New Jersey, and G. A. Mirick, formerly Deputy Commissioner of Education for New Jersey. ^1.25 net. Postpaid. TEACHING LITERATURE IN THE GRAMIWAR GRADES AND HIGH SCHOOL. By Emma M. Bolenius, formerly Instructor in English, Central Commercial and Manual Training High School, Newark, N.J. Jgi.25 net. Postpaid. PUBLIC SCHOOL ADMINISTRATION. By E. P. CuBBERLEV. )f 1.75 net. Postpaid. THE PSYCHOLOGY OF THE COMMON BRANCHES. By F. N. Freeman, Assistant Professor of Educational Psychol- ogy, University of Chicago. S1.25 net. Postpaid. THE MEASUREMENT OF INTELLIGENCE. By L. M. Terman. S1.50 net. Postpaid. Test Material for The Measurement of Intelligence. 50 cents net. Postpaid. EXPERIMENTAL EDUCATION. By F. N. Freeman. 31.30 net. Postpaid. f">lll> I||i""i||l""'i||l> mii"ii|||ii"ii|||ii"n|i|ii"ii|||iiiiii||| t|i|ii«ii||i iji iji i|||i |||i iji ||| |||iiiiir|||iiiiiin|i =-i'l "illlliii'ill'l llllliiiillllliiiillll lIlliiiillllliiiillLiiillllli.iillll illliiiurllliiniillliiiiiilll llliii,iilllli.iiilllii..iillliii,i;lllii„iilllii„iililiii,ulL-J EXPERIMENTAL EDUCATION Laboratory Manual and Typical Results BY FRANK N. FREEMAN, Ph.D. ASSISTANT PROFESSOR OF EDUCATIONAL PSYCHOLOGY THE UNIVERSITY OF CHICAGO AUTHOR OF "THE TEACHING OF HANDWRITING" AND "THE PSYCHOLOGY OF THE COMMON BRANCHES" HOUGHTON MIFFLIN COMPANY BOSTON NEW YORK CHICAGO =Jr =-^iiiiiiil|i i|||i i||i in i||ii"ii||i i||ii'"i||i i|||i"'"i||i""i||i Ill i||iii'ii||ii"'ii||i Ill i||i iii|i i|i""ii|||i"'ii|||i""iij=-= ^iiillllliiiillllliiiilllllri>illlll:iiiillll lllliiiilllllfiiillmiiiillllliiiillllliniillll nil lllliiiilllll lllliiiilllllii»llllliiiillllliiiillllliiiilllliii,iilllii,iii|||n„ii|||ii„| COPYRIGHT, I916, BY FRANK N. FREEMAN ALL RIGHTS RESERVED CAMBRIDGE . MASSACHUSETTS EDITOR'S INTRODUCTION The experimental investigation of educational pro- blems had its origin, in large measure, among the workers in psychological laboratories. The subject- matter of the problems is chiefly psychological in na- ture. Many of the methods which have been devel- oped in the psychological laboratories are applicable to the study of the problems of education, and some of the problems in psychology, particularly those of learning and memory, have direct bearing upon edu- cation. During recent years experimental education has ex- perienced a very rapid growth, and there have been extensive developments in certain specialized direc- tions. The fields in which this work has been largely done are the investigation of the learning process which characterizes progress in the subjects of study in school, the accurate measurement of attainment in these school subjects, and the determination of individual ability through tests of mental maturity, intelligence, and individual differences. These problems, while of the same general nature as those studied in the psy- chological laboratories, are in the main untouched in the work of these laboratories and in the manuals which have been prepared for the guidance of their vi EDITOR'S INTRODUCTION work. Under these conditions the formulation of courses of experiments for students in education, which shall meet the special demands of this particular field of investigation, is highly desirable, since the organization of manuals to serve as guides in such courses has not kept pace with the research which has been carried on. This is the natural result of the new- ness of the field, of its lack of organization, and of the variable quality of the work which has been done in it. The present volume of this series of textbooks is an attempt, by a psychologist who has been engaged in much of this newer type of educational investigation, to make a workable organization of this new field in applied psychology, to sift the valuable phases of the work from those which are ephemeral, and to formu- late a series of experiments which shall introduce the student to the more important problems of experi- mental education. As such it is hoped that it may prove of much use to students of such problems as are here included. Ellwood p. Cubbeblet CONTENTS CHAPTER I. Introduction 1 CHAPTER II. Analysis of Vahious Types of the Learning Process Experiment No.!^ Sensori-Motor Learning .... 13 Experiment No. ^2. Perceptual Learning 25 Experiment No. 3. Learning of the Problem-Solving Type 32 A. Puzzle-Box Experiment 34 X^J- The Tait Labyrinth Puzzle 36 Experiment No.^^. Transfer of Training in Sensori-Motor Learning 41 Experiment No. 6. The Factors in Memory as revealed in Rote Memorizing 51 Experiment No. 6. Memory for Sense Material ... 64 CHAPTER in. Experiments with School Subjects Experiment No. 7. Experimental Psychological Analysis of Handwriting 72 Experiment Noif'S. A Test of Handwriting .... 82 Experiment No. 9. Observation of Eye Movements in Reading 95 Experiment No. lOifStudy of the Perceptual Process in Reading by the Tachistoscopic Method . . . .110 Experiment N<^11. EflSciency of Reading .... 117 Experiment No,M2. Apprehension of Number . . . 132 CHAPTER IV. Tests Experiment No. 13. Tests of Visual Defects .... 141 Experiment No. 14. Tests of Auditory Acuity . . . 152 viii CONTENTS Experiment No. 16. Tests of Maturity of a Mental Func- tion 162 A. Immediate Memory for Numbers .... 162 B. Reconstruction of Sentences 164 Experiment No. 16. Correlation between Tests . . . 170 APPENDIX. Figures and Tests for Use with the Experiments Six Figures for Use with Experiment 2 189 Tait Unicursal Labyrinth, Experiment 3, B . . . . 191 Series of Syllables for Experiment 5 192 Poem to be used in Experiment 6 193 Texts for Use with Experiment 9 198 Figures and Words for Use in Experiment 10 ... . 208 Material for Use with Experiment 12 209 Material for Use with Experiment 15, B 213 Key to Sentence Arrangement in Experiment 15 . . 215 Stimuli for Use with Experiment 16 216 Response List for Experiment 16 217 INDEX 219 ILLUSTRATIONS Fig. 1. Apparatus used in Experiment No. 1 . .14 Fig. 2. Akkangement of Cards used in Experi- ment No. 1 16 Fig. 3. Diagram of Puzzle Box 35 Fig. 4. Tracer Record from a Writer who uses Little (A) and Much (B) Arm Move- ment 76 Fig. S. Tracer Record from a Writer who uses a Medium Amount of Arm Movement . 77 Fig. 6. Mirror for Observing Eye Movements . 95 Fig. 7. Diagram of the Connections for the Tap- ping Apparatus 172 EXPERIMENTAL EDUCATION CHAPTER I INTRODUCTION Within the past ten years there has been very marked activity in the scientific study of the psycho- logical problems involved in education. This activity is evidenced by the large number of papers reporting experimental studies, read before meetings of edu- cational and psychological associations, and by the stream of articles dealing with experimental education appearing in educational and psychological periodicals. This investigation of educational problems by the general methods developed in the psychological lab- oratory was instituted and promoted in the first place by professional psychologists, and originally dealt with problems of general psychological analysis. In the course of its development, however, both the aim and the content of this study have shifted, and it has be- come necessary to reformulate them in order to meet the demands of students who are interested in the psychological aspects of education. The earlier applications of psychology to education dealt primarily with the development of general psy- chological principles, and secondarily with their appli- cation. The first consequence of this emphasis on the 2 EXPERIMENTAL EDUCATION theoretical side was that, since there were no clear standards as to what was of most importance from the practical point of view, much time was spent on the elaboration of principles which were relatively miim- portant for education. The second consequence was that, since the basis of classification was the psycho- logical system, the facts or principles which applied to any particular problem, such as reading or writing, were scattered through various parts of a course in- stead of being focused on the practical problem itself. The course which is here presented accordingly pro- ceeds, for the most part, from the opposite starting- point. Instead of developing the general psychological principles and laws, and then making incidental appK- cation, it attacks directly the practical problems and attempts to throw light upon them by an analysis of the psychological principles which are involved in them. This means that the selection of problems is based directly on their practical importance, and that the arrangement is such as to bring together the facts which serve to illuminate some one practical problem. An exception to this general rule that the experi- ments deal with specific educational problems rather than with general psychological principles, appears in the early experiments, which deal with learning and memorizing. This exception is justified by the very wide and evident application which these principles have in education. The purpose of the experiments in this first section is to cause the student to make an INTRODUCTION S analysis of his own learning process, and to compare his learning process with that of others, to the end that he may know what the conditions of eflBcient learning are, and how the learning of others may be most effec- tively directed. For this reason the conditions of learning are experimentally varied in such a manner that the effect of different conditions may be compared. A variety of typical forms of learning are employed, including rote and sense memorizing. In the next group of experiments the student is led to examine forms of learning which are represented in the fundamental subjects of the school curriculum. We are here dealing with special problems which have a stiU more direct application to school procedure than does the analysis of the learning process in general. Besides the analysis of the mental process involved in the learning of the school subjects, it is of great value to the administration of education to be able to meas- ure the attainment of pupils in the various branches by means of tests which render the results comparable to general standards. Accordingly, two examples of such test methods are included in the experiments on the school subjects. Finally, for the proper direction of the pupil's learn- ing process, it is necessary that we shall have informa- tion regarding his intellectual equipment for the tasks which are required of him. The detection of defects in the avenues by which experience is acquired necessi- tates the discovery and correction of, or allowance for, 4 EXPERIMENTAL EDUCATION sensory defects. A general knowledge of the methods of testing for sensory defects is desirable. The methods of testing other types of ability can be illustrated only roughly, and the qualification must always be kept in mind that the results with adults cannot be applied directly to children. Various test methods can be illus- trated, however, so as to give an acquaintance with the types of tests which give an insight into important individual and age variations. Opportunity may also be given to become familiar with the methods of study- ing the interrelation of mental traits and with some of the typical results of such a study. , ' In addition to an appreciation of the facts of learn- ing established by the experiments, the course should stimulate the student to make a study of the process of experimentation itself, and to form a notion of the chief characteristics of a valid experimental method. The main features of the experimental method have been frequently described, and may be briefly illustrated here as an introduction to the experiments themselves. The aim of scientific investigation may be said to be to study the uniformities in the facts or processes which are the subjects of study. This usually resolves itself into a study of the relationship of one series of events to another as to cause and effect. Thus, the study of the events of the physical world has led to the discoveries of certain xmiformities or laws, such as the law of gravitation, the laws of light and soimd propa- gation, and the law of evolution. Now, the discovery INTRODUCTION 5 of such uniformities merely by the observation of the events as they occur spontaneously, is very difficult, since a particular event may never occur twice with precisely the same surrounding conditions, or two events between which the relationship is to be discov- ered may always occur intermingled with a variety of others. Thus, it was difficult to determine purely by observation whether or not the bite of a mosquito is related to yellow fever as cause to effect, because there were associated with this condition a variety of others — such as contact with those who were afflicted with the disease — to complicate the relationship. Under these circumstances the experimental method was re- sorted to in order to simplify the conditions by artifi- cially producing the events which were to be studied, in isolation from others. That is, an investigator first ex- posed himself to contact with a yellow- fever patient while carefully protecting himself from the bite of the mosquito; and then exposed himself to the mosquito bite while remaining isolated from infected patients. The experiment proved to be crucial and conclusive by the immunity of the investigator in the first case and his infection in the second. Sometimes the aim of an experiment is not so specific as this. Instead of seeking to determine which of two alternative explanations is correct, the experimenter may attempt to make a more general analysis or survey of certain phenomena, for the sake of discovering any laws which may be revealed, and of hitting upon prob- 6 EXPERIMENTAL EDUCATION lems which may be attacked in a more specific manner afterwards. Such an experiment is illustrated in the well-known Bryan and Harter experiment in learning the telegraphic language. In the first survey experi- ment the authors discovered the existence of a plateau or long level place in the ciu-ve of progress. Their later experiment was for the purpose of discovering the cause of this phenomenon by further analysis of the learning curve into several curves, representing prog- ress in receiving isolated letters, disconnected words, and connected discourse. Whether an experiment is of the crucial or of the general-survey type, it has certain characteristics which must be adhered to if it is to give results which are reliable. First, all the conditions which might affect the result in any degree must be controlled, as far as possible; or, if they cannot be controlled com- pletely, they must be accurately measured or observed and recorded. The simplest method, and the one which is pursued as far as possible, is to keep all the conditions constant except one, and then to vary one condition by a certain known amount, and measure the result. Thus, we may measure the effect of practice by giving a person a certain specified amount of practice in a particular activity, and by noting the change in his ability without introducing or allowing a change in any other condition which would affect the result. If the person were in better health in one part of the period than in another, or if he also practiced in some related INTRODUCTION 7 field which afifected his progress in the one under inves- tigation, these factors would have to be taken into account. It is frequently impossible or very difficult to study one factor in isolation, and in such cases it may be necessary to make check experiments. J. E. W. Wallin made an investigation * of the effect of correcting de- fects of the teeth and instituting correct habits of care of the mouth, in the case of school children, which illus- trates this point. A group of children were tested men- tally before, during, and after the correction of mouth defects and the institution of habits of hygiene, which tests continued for several months, and a large im- provement was found. But other factors than oral hygiene affected the results; namely, general increase in maturity, school work, and the practice with the tests themselves. Wallin himself recognized this fact, and wrote that it had been his plan to give check tests to other children who had not received the special treatment. This would have made it possible to dis- count the other factors. This principle of the check test has a very wide application, and will be illustrated in several of the experiments of this course. Whatever the special character of an experiment, the procediu'e must be objective and verifiable. The possi- bility of verification by the repetition of the experi- ment by another rests upon the objectivity of the 1 J. E. W. Wallin, "Experimental Oral Euthenics"; in Dental Cosmos, April and May, 191S8. 8 EXPERIMENTAL EDUCATION procedure. It therefore is necessary to be clear as to the meaning of objectivity. Objectivity is contrasted with unsupported opinion. In a scientific experiment it requires that the conditions be so controlled and reported that they may be duplicated. It means fur- ther that the results be carefully measured and pre- sented, or, when measurement is not possible, that they be fully and accurately described. It means finally that the conclusions which are drawn shall rest upon the results which are presented, so that another person may check them up by reference to the results on which they are based. These are, in brief outline, the chief requirements of a valid scientific experiment. A word should be said about the presentation of results. When a number of experiments are made with the same subject, or when a number of subjects are experimented upon, it be- comes necessary to employ some method of formulat- ing the mass of results so as to show their trend. Fre- quently the significance of the results is dependent upon the skill with which this formulation is made. Two general methods of formulation may be distin- guished; the tabular presentation or statistical formu- lation of numerical data, and the graphic presentation of data. Both of these methods will be illustrated in the presentation of typical results of the following ex- periments. A concise account of the chief methods may be found in G. M. Whipple's Manual of Mental and Physical Tests, vol. i, chap. 3. INTRODUCTION 9 The chief conditions for a valid experiment may be smnmed up in a set of rules: — 1. So far as possible a single factor should be iso- lated for study. When this cannot be done, check experiments should be made. 2. All the conditions of the experiment should be made quantitatively as precise as possible. This includes the regulation of the stimulus as to the amount, duration, and interval, and the meas- urement of response. Where required, it also involves the correct statistical formulation of results. It should go without saying, but unfor- tvmately does not, that all calculations must be accurate. Every calculation should be checked. 3. In some cases a qualitative interpretation gives the results more significantly than a quantitative formulation of results. 4. The analysis or interpretation of results is ex- tremely important, particularly in a survey experiment. 5. An immediate record should be made of every- thing which occurs, as introspections, and all records should be completely identified. 6. An objective, impartial attitude toward the re- sults is essential. 7. The problems which are raised by the experiment should be noted as well as the solutions which are obtained, and, if possible, modifications of the experiment should be set up to attack them. 10 EXPERIMENTAL EDUCATION 8. The conclusions should be generaUzed only so far as the results justify. The value of an experiment to others depends on the way it is reported, as well as on the character of the experiment. To make the report complete it should include : — 1. A brief but clear statement of the problem. 2. An account of previous investigations leading up to the present one. 3. A more detailed statement of the problem and method of the experiment, including a sketch of the apparatus. 4. A presentation by tables, graphic figures, and explanatory statements of the detailed and simi- mary results. 5. An interpretative discussion of the results. 6. A discriminating statement of conclusions. 7. A bibliography of the references used in the report. In the set experiments of this course Nos. 2 and 7 may be omitted, but they are necessary in an original investigation. Note to the instructor Most of the experiments which are here outlined are intended to be performed by pairs of students working together. Each student should work both as experi- menter and as subject. Each student is expected to turn in a full written INTRODUCTION 11 report, which should include the numerical or other data which were obtained by himself as subject and a complete analysis and interpretation of the results. After all the individual reports are in for an experi- ment, some member of the class should combine the data from all of these into a group report. This group report should be presented at a class meeting, and thoroughly discussed. This presentation and discus- sion of the group results is the most valuable phase of the whole work. It makes it possible to discover the general results as they emerge from the variety of the individual results, and also makes each experiment a study of individual differences. Furthermore, it gives excellent opportvmity to comment on violations of correct experimental procedure and to bring home the principles of vahd experimentation. Note. — The apparatus used in this course can be ob- tained from the psychology shop of the University of Chi- cago by writing to the author, or from C. H. Stoelting & Co., Chicago. Prices will be quoted on application. CHAPTER II ANALYSIS OF VARIOUS TYPES OF THE LEARNING PROCESS One of the most fruitful fields of experimentation in education is the analysis of the learning process. Many experiments in this field have been made, and a variety of types of learning have been studied. The types of learning which have been the subjects of investigation may be classified, roughly, into four groups: — A. The development of a motor coordination. B. The development of adequate perception. C. The formation of associations between per- ceptual or ideational elements. D. The analysis of a situation. Such a classification as this is of value because it distinguishes between processes which differ from the descriptive point of view, as well as in the methods which are best suited to hasten progress in the several kinds of learning. It must not be inferred, however, that the same task may not include more than one kind of learning, or that similarities between the va- rious kinds of learning cannot be found. THE LEARNING PROCESS 13 EXPEBIMENT No. 1 SENSORI-MOTOE LEARNING Problem. The first experiment is for the purpose of illustrating the development of a motor coordination. The development of motor coordination in its simplest terms consists in the adjustment of a movement to a stimulus, or, more particularly, in the development of the ability to make movements which shall meet cer- tain conditions presented in perceptual form. Physio- logically, motor learning consists in the formation of appropriate nervous connections, or paths of discharge, between certain sensory centers and certain motor centers. This is the matter reduced to simplest terms. The process of such development is illustrated by J. H. Bair's experiment in learning to move the ears. There being no instinctive connection between the sensations which accompany ear movements and the movements themselves, this connection has to be formed through practice. A more complex type of learning consists in adapting a series of movements to a complex stimulus or to a changing stimulus. Swift's ball-tossing experi- ment, and the experiment included in this course, are relatively simple examples of this type. The matter becomes more complex when it becomes necessary to organize the perceptual or motor elements, as in Bryan and Barter's experiment with telegraphy or Swift's or Book's experiments in learning to use the typewriter. 14 EXPERIMENTAL EDUCATION Such experiments have a direct bearing upon educa- tion wherever it is concerned with the development of motor skill, as in handwriting, drawing, and the man- ual arts generally. The results of such experiments also have an indirect bearing on education by means of the discovery, through them, of general principles of Pio. 1. DRAWING OP APPARATUS USED IN EXPERIMENT NO. 1, SHOWING THE GENERAL EELATIONS Olf THE MECHANISM the learning process — as, for example, the relation of effort to progress in learning. Material and method. The learning process in this expteriment consists in adapting simple hand move- ments to the drawing of a series of lines to connect a group of dots, when the relation between the direction of the hand movements and of the movement of the pen which produces the lines is different from that to which one is accustomed. The modification in the rela- tionship between the hand movements and the result- THE LEARNING PROCESS 15 ing pen movements is made by means of a mechanism, the study of which is left to the student. (See Fig. 1.) In addition to the modification produced by this mech- anism, the apparent direction of the pen movement is changed, not by mechanical adjustments, but by the use of a mirror. By varying the position of the mirror, this apparent direction may be changed at will. For the study of progress in learning, let one person of each pair place the mirror parallel to the long side of the board which is farther from the subject. Let the sec- ond member of each pair place the mirror parallel to the left side of the base. The apparatus should be placed so that the long side of the base is parallel to the edge of the table, and the handle in a convenient position to be manipulated by the right hand. Stimulus card No. 1 is to be used in this experiment, and is to be placed with the margins parallel to the edges of the apparatus and with the X toward the subject. The stimulus cards are shown in the accom- panying Figure 2. To indicate the order in which the lines are to be drawn, small groups of dots are used, instead of figures, since the figures would have to be reversed in order to be read in a mirror. The lines are to be drawn from circle 1 to circle 2, then to circle 3, and so on; and the star should be finally completed by connect- ing circles 6 and 1. The aim of the subject should be to cause the pen to touch each dot in succession 16 EXPERIMENTAL EDUCATION • •• O • o • < o •. o •• o • • • • • • Stimulas Card No. I. O • • o '•• o o •• • o stimulus Card No. n. Fio. 2. ARRANGEMENT OF CARDS USED m EXPERIMENT NO. 1 as rapidly as pos- sible, and the cri- terion of progress may be the reduc- tion of the time taken to complete the whole figure. Since the accu- racy requirement is constant, name- ly, that each dot shall be touched in succession, the time taken serves as a convenient and sufficiently adequate measure of efficiency. The experi- menter should note the time of each trial with a stop - watch, and should attend to changing the cards, keeping the pen supplied with ink, etc. Twenty-five tri- THE LEARNING PROCESS 17 als shovJd be made with the card and mirror, as indi- cated above. The subject should make note during the experiment of any introspections he may make regarding the means by which improvement is attained, the direc- tion of his attention, etc. In order to throw light upon the value and Hmita- tions of instruction in this type of learning, the follow- ing method should be employed: Let one student of each pair first go through the experiment without in- struction or any special consideration of the methods of learning. Let him then develop, as best he can on the basis of his own experience and of his reading, the principles and rules of economical sensori-motor learning, and instruct his partner in these rules and principles. These instructions should be formulated in writing and included in the report. They should not be made too specific, but should be of such a kind that they could be applied to other tasks in learning which are of a similar nature. The second student may also obtain what light he can on the best method of procedure from an observation of his predecessor, and should give an accoimt of his conclusions in his report. He should also discuss the question whether the instructions were of assistance to him. The gen- eral report should include a comparison of the records of the instructed and uninstructed learners. Treatment of results. A curve of progress should be constructed by charting, on cross-section paper, the 18 EXPERIMENTAL EDUCATION time taken for the successive trials. Each trial may be represented by a imit on the horizontal axis, and the height of the curve above these successive unit dis- tances may represent the length of time required for the succeeding trials. Thus a drop in the curve means decrease in the time required, or improvement. An analysis of the practice curve should be made, so as to explain its general form and any fluctuations which may appear. In the analysis and interpretation of the curves, use should be made of the notes on introspections. The numerical results which are the basis of the practice curves should be presented in the form of tables. The general report upon this experiment should discuss, first, individual differences with reference to the rate of improvement, the amount of skill attained, the suddenness or gradualness of improvement; and, second, the general facts common to the individual results, as the part of the practice series at which the more rapid part of the improvement occurs, and the effect of intelligent instruction upon learning. A gen- eral summary shoidd also be made of the individual analyses and interpretations. This should include a discussion of the difficulties which had to be overcome and the methods and devices which were employed in overcoming them. Results of the experiment. Some of the most signifi- cant results of this experiment are derived from an THE LEARNING PROCESS 19 analysis by each subject of his experience in the learn- ing, and from the interpretation of this experience so as to apply the conclusions to similar forms of learning, such as handwriting. In order to give opportunity for some originality in this analysis and interpretation, the discussion of these matters will not be entered upon here. The province of " instruction " in this type of learning, and the kind of instruction which is of most value, is directly related to the learning process. Typical objective data on which the interpretative discussion just mentioned may be based are presented in Chart I. It is clear that there is some radical differ- ence between the " instructed " and the " uninstructed " groups which is to be explained. The difference is not accidental, since this is a typical case. The ques- tion should be raised whether the verbal instructions in this case constituted the only factor which was dif- ferent in the two groups. If the analysis of the situa- tion should lead to the conclusion that there are other differences between the groups introduced by the con- ditions of the experiment than that of the presence or absence of verbal instructions, the importance of these other factors should be estimated. If there is opportu- nity to carry on other experiments the relative impor- tance of instruction and of these other factors may be measured. The median curves give a good basis for the discus- sion of the general form of the practice curve, and for an attempt to explain it. In this discussion it is well to I I 1 r t fm I I ^ t T 60 ^ V t A I ^ 40 j r K^ \ ^Z% S 1 ^-^ ^ , — v. ^ ~- ^ ■"- Trials 10 15 20 26 Chabt I. CURVES OF PROGRESS IN MODIFIED MIR- ROR DRAWING BASED ON THE MEDIANS OF TWO GROUPS OF INDIVIDUALS The upper line represents the " uninstructed " group (eleven individuals) and the lower line the " instructed " (nine indi- viduals). Time in secondB u 200 \ \ 180 2 140 120 100 \ 80 V \ 1 • 60 \ . 1 / ^ ' ' V 1 L \ / \ \ / \ -^ \ \ ^ \ ^ s ' \ \ \ 20 r ' s \ \ \ \ y X V s ■s pj ^ — — 1 ~-' - N ^ . _ 10 20 2 1 3 25 Chabt n. CURVES OF PROGRESS OF THREE INDI- VIDUALS IN MODIFIED MIRROR DRAWING Showing individual differences in rapidity of the sensori- motor process and in rate of improvement. Tlie fluctuations in an individual's practice curve are also brought out. It should be noted tliat the scale of this chart in respect to the represen- tation of the time of the trials is reduced, making the height of the curves half that of the curves of Chart I. 22 EXPERIMENTAL EDUCATION compare the form of these curves with that of others, as, for example, those of Bryan and Barter's experi- ment, and the ball-tossing experiment of Swift. The great individual differences which appear in such a form of learning as this are clearly illustrated in Chart II. These extremes are not at all unusual. Their significance and educational bearing may be discussed. The introspective notes should throw some hght on the fluctuations which are to be observed in the individual curves. A comparison of the records of those individuals who attempt to analyze the relationship of the hand move- ment to the pen movement, in order to guide their efforts by a comprehension of this relationship, indi- cates that, on the average, they take more time to learn than those who abandon such attempts and resort to random trial. In one group the average time spent by those who attempted analysis was forty-five seconds per trial, and by those who did not, thirty-one seconds. Extension of the experiment. It would be worth while to know whether those who make a thorough analysis before starting are benefited thereby in their first set of trials, and also whether they do better when the conditions are modified, as in Experiment No. 4. To study this question a class may be divided into two equal groups. One group may then proceed according to the regular directions, and the other may remove the cover, study the levers, and work out by diagram THE LEARNING PROCESS 23 the relationship of the pen and the hand movements to the movements of the hand. This should be done without practicing making the movements with the hand. A marked plateau does not usually appear in this particular form of learning so far as it is carried. The experiment may be extended by adding other trials to determine whether a plateau would then appear. A further extension may be made by trying the experi- ment with children of different ages. If opportu- nity offers, the correlation between ability in this task and in other forms of sensory motor learning might be found. EXPERIMENTAL EDUCATION QUESTIONS AND TOPICS FOR DISCUSSION These questions are intended to stimulate reflection upon the wider applications of the experiment. 1. What are the likenesses or differences between the problem confronting the subject in this experiment and the problem before the child in learning to write? 2. What is indicated by the experiment regarding the sphere of the trial and success method in learning? Of what value is the attempt to analyze the conditions of the problem? 3. a. What bearing does one's general attitude, as of confidence or the reverse, have upon progress? h. What bearing does physical condition have? c. What is the effect of special effort? 4. Does a knowledge of the laws of learning have any beneficial effect upon the control of the feelings and one's attitude toward them? 5. In what respect and to what extent is instruction of value in facilitating progress? 6. What bearing do individual differences have on the develop- ment and application of instructions? 7. Is the possibility of analysis greater or less in this experiment than in the child's sensori-motor learning? 8. What is the role of repetition in such learning as this? yQ. How are trials made more correct — by a study of the move- ^ ment or of the external result? 10. What place has the study of the movement? REFERENCES Bair, J. H. "The Development of Voluntary Control"; in Psy- chological Review, vol. vni (1901), p. 474. Bryan and Harter. " Experiments in Learning the Telegraphic Language"; in Psychological Review (1897), pp. 27-53; (1899), pp. 345-575. Book, W. F. Psychology of SkiU. Dearborn, W. F. "Experiments in Learning"; in Journal of Educational Psychology, vol. i (1909-10), p. 373. Ruger, H. A. Psychology of Efficiency. Starch, D. Experiments in Educational Psychology, p. 34. Swift, E. J. Mind in the Making. THE LEARNING PROCESS 25 Experiment No. 2 perceptual learning Problem. The second experiment deals with a type of learning in which the perceptual element is pre- dominant. This element enters into many forms of learning which seem at first sight to be mainly motor in character. In handwriting, for example, the per- ception of form precedes its reproduction, and also is made more precise by the effort at reproduction. In typewriting the comprehension of the relation of the letter positions on the keyboard to the letters which compose a printed or imagined word is at first chiefly a matter of perception. Finally, it is now recognized that in drawing the perceptual element is of more im- portance than the motor element. The ability to ar- range proportions and directions of lines in such a way that they will constitute the picture of an object is [ more a matter of seeing relationships in space properly than of skill in handling a pencil. Drawing a figure in which the elements present no technical difficulty is therefore an excellent means of testing the accuracy of perception and of tracing the development of per- ception. Material and method. For this purpose a series of six figures may be used (and these are reproduced in the Appendix) . The experimenter should be prepared with a watch having a second hand, and a series of W EXPEBDIEXTAL EDUCATION cards upon which the figures are drawn. Tte cards should be placed face down and arranged in order, with the first one on top. The subject should be pro- vided with a number of sheets of paper the same size as the cards. After a ready signal, the esperimejiter should hold the first card in sach a posttson that the subject may see it dearly for ten seconds. When the card is lowered, the subject should immediatelj' draw what he can of the figure and then turn or fold the sheet ovCT so as to conceal what he has drawn. The figure shoTild then be presented again and drawn from memory. This procedure may be repeated until the subject is confident that he has mastered the figure. The experimenter should observe and note any actions on the part ^ n «K;uii)a;sj,9«tuaafi- otewsioooowwciooooociujo »«SOT«-iflCM«»Oa»lOCN«SO(H*CD BuiujiDap^ JOO ; --(II j ■■»<» 1^ j j ; ^(O ■ • 90U9d9ffl(l :7 : : I : :+ 1 1 1 : : : ;7 : : oi 1 Bmujbvsq .(^1 . -^ . .U3a - - - -o --i^es ■ '. CO duivMosq i:::: -"S ::•:"•:": ; s 90Ugjt9ff}Q : :7+ ; : 1 : I +++7 : :7 :+ l" ■papruisuj S :SS : > isSSSgS : :3 :S GO 90U9J,3jfl(J :7 : : 1+ :+ : : : : : 1+ :+ : 00 l' •psptupmuQ ^« ■ -INO ---J t-OO -00 • a la 02 pns}A to s ^ n fUojipny Oi—SDWWt-OnCOOSt-i-HUJCOOIOOi-lCSi-l (N CO a* i-< fH »0 CH i-< rH <« i-( i-H ,-1 ,-1 en ot »o 1 1 <»uQc^CMdni-it-^Mi4^^'dp^c^P^ 1 > < 58 EXPERIMENTAL EDUCATION TABLE IV. NUMBER OF PRESENTATIONS NECESSARY FOR EACH INDIVIDUAL OF A SECOND GROUP IN THREE SERIES Individual Series I and la Series II Series III Audi- tory Visual Unin- struded Difference In. structed Differ- ence A B C D E F G H I 13 15 16 21 22 23 23 24 28 32 40 49 61 78 8 10 18 18 16 10 27 14 18 17 34 17 48 36 9 14 9 11 13 2i is 33 + 1 + 4 - 9 - 7 - i + 'i -'i -15 is 8 6 28 20 ig -i - 2 -21 +i6 J K L M N -14 -17. Average... 31.8 20.8 15.6 -3.1 16.0 -7.5 Results of the experiment. Specimen results from this experiment are shown in Tables III and IV. Not much reliance can be placed on some of the results, since some of the members of the class failed to follow instructions carefully. There are some fairly clear out- standing facts, however. By reference to both groups it appears that the series were learned more quickly by visual than by auditory presentation by the majority of the subjects. Only individuals C, D, E, I, and M of Table III, and C and THE LEARNING PROCESS 59 G of Table IV, learned more rapidly by auditory presentation. While the visual presentation proved to be the better on the average and for the large major- ity of the cases, the exceptional individuals must not be overlooked nor looked upon as abnormal. Apparently the instructions did not have a marked effect upon the rapidity of learning. While the in- structed group of Table IV gained over twice as much in Series III over Series II as did the uninstructed group, in Table III there was no appreciable gain by either group. This may be due to several reasons. There was some indication that the uninstructed group hit upon some of the devices mentioned in the instruc- tions independently. The extent of individual and ac- cidental differences may have covered up real differ- ences due to instructions. The instructions may not have been of the best. More prolonged practice may be necessary to profit by instructions. These possible explanations should be experimentally tested before it is concluded that instructions are of negligible impor- tance. This furnishes problems for further extension of this experiment. Learning beyond the threshold evidently produces greater permanence of learning, as is to be expected. The extent to which learning should be carried beyond the threshold is evidently to be determined by the purpose in learning, — by the degree of permanence that is desired. While attempting to recall during learning did not. 60 EXPERIMENTAL EDUCATION according to the figures of Table III, lower the number of presentations necessary in comparison with continu- ous presentation, it did produce somewhat greater permanence. The results from other groups commonly come out in favor of the method of attempting to recall at intervals, both in the number of presentations necessary for first learning and for relearning. The averages in the last two columns of Table III bring to light individual differences in rapidity of first learning and of relearning, and in the correlation be- tween rapidity of learning and permanence. In order to make the inspection of the table easier the scores may be arranged in ascending order. This is done in Table V. Very large individual differences appear both in the first learning and the relearning, the ratio of the lowest to the highest score being about 1 to 6 and 1 to 7 respectively. This is a very large difference among individuals of a group who are rather homog- enous in training and general ability. The calculation of a correlation coeflBcient is not in this case a very profitable proceeding as a means of measuring the degree of correlation, since the group includes several classes of cases, each characterized by a different relationship of speed of learning to reten- tion. As a very rough method of indicating the degree of correlation in general, however, we may average the relearning scores of the top, middle, and bottom thirds of the whole group, classified on the basis of the first learning scores. The averages of the relearning scores THE LEARNING PROCESS 61 TABLE V. SCORES IN FIRST LEARNING AND IN RELEARNING First learning Releaming A 6.8 9.8 11.0 11.4 11.8 12.2 12.7 13.6 15.2 15.6 17.0 18.6 19.0 20.2 25.8 31.0 34.2 37.6 2.0 2.5 3.0 3.0 4.0 5.0 4.0 2.5 1.5 5.0 5.5 6.0 2.0 10.0 3.5 2.0 5.5 4.0 B C D ^ E. .. . F Average, 3.25 G H I J K L Average, 4.1 M N P Q R Average, 4.5 show a slight tendency to increase with the increase in the first-learning scores, indicating a very slight ten- dency for the learning of rapid learners to be more retentive than that of slow learners — so far as this test goes. But the slightness of the correlation is shown by the fact that if the highest releaming scores of the middle and highest thirds were exchanged, namely, the 6 and 10, the averages of these groups would be 4.7 and 3.5 respectively. A much more significant mode of examining such an array as this is to note the kind of cases which are rep- 62 EXPERIMENTAL EDUCATION resented. It is apparent, after a moment's inspection, that there are those of high, medium, and low reten- tiveness among the fast, the medium, and the slow learners. We cannot accept the traditional view that slow learners are more apt to be retentive than rapid learners. So far as there is a general rule the latter seems to hold. But no one type of correspondence appears to be general. Extension of the experiment. This experiment may be extended by varying the conditions and the material in a variety of ways. The effect of instructions may be studied more fully, as already suggested. The effect of learning beyond the threshold on permanence, and the effect of attempting to recall on both speed and perma- nence may be studied by having the same individual use the various methods with different series. All the different devices may be tested with series of words, or with vocabulary pairs instead of with nonsense sylla- bles. In such case the variability in the difficulty of material will be much greater. The effect of distribu- tion of the repetitions may be studied by taking series of the same length or greater length, and making a few repetitions of each series at a sitting instead of carrying them to the threshold at one sitting. The effect of distribution on permanence is particularly important. THE LEARNING PROCESS QUESTIONS AND TOPICS FOR DISCUSSION 1. Are any parts of the seriefe learned before the rest? If so, ac- count for the fact. 2. What accessory devices, if any, are used to aid learning? Do the stimuli arouse any form of imagery? If so, what form or forms? What function does the imagery have, if any, in the learning process? 3. Is it correct to assume that the imagery employed belongs ex- clusively to the sense through which the stimulus was given? 4. Is there a definite point at which memorizing may be said to be absolute or complete? If so, how may it be described? If not, how may degrees of learning be described? 5. Can the effect of learning in the early stages always be meas- ured? Why does command of the material sometimes seem to deteriorate in the early stages? 6. An examination of recruits in Germany showed that they could recall little of what they learned in school. How is this result to be interpreted? 7. Give illustrations of the value of learning of various degrees of completeness. 8. Draw conclusions regarding the various factors or methods studied. 9. What is the effect of the formation of wrong associations? 10. Is rote memorizing in the school justified? Under what cir- cumstances? REFERENCES Colvin, S. S. The Learning Process, chap. xi. Ebbinghaus, H. Memory. Tr. by Ruger. Teachers College Pub- lications. Lange, R. Praktisches Handbuch fiir den Rechtschreibunterricht. Lay, W. Fiihrer durch den Rechtschreibunterricht. Meumann, E. Psychology of Learning and Varlesungen zur Ein- fuhrung in die ExperimentaUe Pddogojik, vol. ii. Whipple, G. M. Manual of Mental and Physical Tests, p. 356. 64 EXPERIMENTAl- EDUCATION Experiment No. 6 memory for sense material Problem. The problem of this experiment is to find the best method of memorizing sense material. Some of the same alternative methods tested in the experi- ment on rote memory might also be tried with logical memory. For example, one could determine the value of attempting to repeat the selection which was being learned before the memorizing was complete, or one might study the effect of memorizing beyond the threshold. In this experiment, however, but one phase of the problem will be attacked. The aim will be to determine whether it is better to memorize by the so- called part method, or the whole method. In using the part method, the learner memorizes a sentence or stanza or other small part at a time. In using the whole method one memorizes by reading the whole selection, or a fairly large part of the selection at one time. In- stead of using the whole method in the strict sense, it will be more useful to modify it by dwelling somewhat longer upon the more difficult parts of the passage after they have been discovered. This is sometimes called the combined method. Material and method. In this experiment each sub- ject may work alone. While the memorizing need not be done in the laboratory, it should be done during the laboratory period in order to have the time of day uni- THE LEARNING PROCESS 65 form. In order that the material which is used by the class may be miiform, copies of a poem have been prepared. The selection is printed in the Appendix. In order to compare the two methods in question, the simple procedure which is likely to suggest itself at first thought will not suffice. This method would be to spend first a certain amount of time in memorizing with one method, and then an equal amount of time using the second method, and compare the amount learned by the two methods. The difficulty with this procedure is that there is a decided improvement in memorizing with practice, which operates to the ad- vantage of the second method used. This practice effect may be in a measure offset by using first method number 1, then number 2, and finally number 1, and averaging the rates of learning during the first and third periods by method number 1. Then, to make sure that the procedure is fair to both methods, still another precaution may be taken; viz., to let one half the class use the part method first and the other half the whole method. In the following plan, let A repre- sent one of each pair of students working together, and B the other: — A B Part method — 40 minutes Whole method — 40 minutes Whole method — 80 minutes Part method — 80 minutes Part method — 40 minutes Whole method — 40 minutes The work should be done in forty-minute periods, and should be distributed over two days. The whole 66 EXPEEIMENTAL EDUCATION poem should be read once by all to get the general course of thought. The material which is being memorized should be studied until it can be repeated once without error, at least so far as the learner can tell. When the part method is used, work should be continued until all the parts can be repeated continuously, and not merely as separate parts. If the material which is being studied is not quite finished at the end of the allotted time, the period may be extended. The efficiency of the methods is to be measured by the number of lines per hour which can be learned by their use. The subject should estimate, when using the whole method, the number of lines he can learn in the allotted time; but, if this amount is not learned, he should continue either then or on the next day until it is learned. If considerable time remains, he should learn another section. In order to compare the efficacy of the two methods, with reference to permanence, the parts learned should be relearned at the end of a week and the amoimt of saving effected should be tabulated. Each part should be relearned by the same method that was used in first learning. Treatment of results. Each student should describe in his report the order in which he used the methods, the amount of time devoted to study by each method, and should calculate and report the number of lines per hour learned by each method. The amount of im- provement, if any, from the first to third period should THE LEARNING PROCESS 67 be calculated. The topics and questions given below should also be discussed. In the general report, a table should be given which shows the lines learned per hour by each student in the various periods and by the two methods. Those who begin with the same method should be grouped to- gether. The averages necessary to bring out the com- parison of the two methods should then be calculated. The practice effect should also be calculated. A com- parison of the degree of retention in the case of the fast and of the slow learners may be made. Results of the experiment. The tabulated results of this experiment for one section of nineteen are given in Table VI. After the somewhat detailed analysis of some of the preceding experiments, the student may be left with a few indications of the outstanding facts in this table. He may then elaborate the details. It appears, contrary to the principle which is gen- erally accepted, that the part method gives better results on the average, and in the case of the majority of the individuals, than does the whole method. In some sections the whole method gives better results on the average, but there are still some individuals who do better with the part method. This so far as first learning is concerned. In the case of permanence of memory, as we should expect, the whole method makes a better showing, though even here some individuals do better by the part method. Let the student weigh the statement that we should expect the whole method cs y. »— 1 N « O S S H ■g pel a fe. g oo»ao«a»»oos»«« »« S eot-'*«5io«*-'*(O-!f.60i»OS'« »o i w ^ i? : :o : ioe»i-iuio5 ■«* ■i >g; ^■5 • . ■* . 00« S Bi 1 S± Cb «s l> »^ o.= ffliooooiojOTWeo 1> wt-'«ftO"* = 1? • :iN :ooooi>; •< • . t- . ooco«5»m w •^ /^ .-i&»(«ar-« « o 05 s 5. e MS t- i? oo(o.-iQOO»«Jiflao •* «■& cctowe©-*"*-*— !•* •o 1^ '"' "o ;3 ■13 S S) 1 g •^ HO^J^lddpicri < n I: g .5 i» -"^ .1 n :S§S5§ : 'i : 1 1 : :^ c? "e'w bi §0 -SSooOJt-rt o J o ;OTq-*»i»OrHus *^ S-S fc ^ M a 1 1 ^ ■i :b-«CDOCD * 'O ^ •« .(DtNCOCOOO ■ ■{« : i 1 i ^ ;o»(« w^m : :« i ■e ; o o CO o »« : " eo ! 00 5 ;Sj55°^S;S : ;§ '. o* 1 «S US «1 us § g ,-H » CO m o o ws w w j>! •«' tZ (M 00 -* w t« -ji « m Tfi 00 OT 1 <« 1 ■e "O u » ^ t-osooeocooooioifo to ^ OTGO«flmt«mQ OTCO«D(K'*GOi-H-*'*0r CO K g? '|>CDCDO(D ; "O CO f.! .»gooooco ; ;o* M •<£ ■s'" Q* ^ 1^ fe 1 Eg lo «s «: CD ^ B.H .-HOCDCo'oO'flCl'-it- ■* ^•- (Neo-^o^m-^fiMWTfi©- CO it 'O 'W500 : 'O ;tji ;oo ws . .eo t- rl "E »^ (^ (a. <5 to , _C »o ^ §,e SOOCOr-lOOCTeOOW ■* CJOO'^CCOOOQOTOTBOcr CO "a 9 ra :s % 1 •S A 1 .5 < p: ij c ^ !i: HH >i p- a ' -S 70 EXPERIMENTAL EDUCATION to be superior for permanence of memory, and argue for or against it. It is possible that previous practice with the part method outweighs the inherent advantages of the whole method. This can be tested by inquiring into the previous habits of memorizing of the individuals who made better time by the whole and the part methods respectively; and by having some individuals, who do better by the part method, carry on more ex- tended practice with the whole method to find out whether the advantage is reversed. The correlation between speed of first memorizing and permanence of memory may be examined, as in the experiment with nonsense syllables. In the group under examination the same rule holds. While, in general, the rapid learners relearn more quickly, some of them relearn slowly and some of the slow learners relearn quickly. Extension of the experiment. This experiment may be extended by carrying on more prolonged tests with a variety of materials; by experimenting with different amounts which are taken as sections to be learned as a whole — particularly with reference to the individual- ity of the learner and the difficulty of the subject mat- ter; and by making the test with children of different ages under school conditions. THE LEARNING PROCESS 71 QUESTIONS AND TOPICS FOR DISCUSSION 1. What, in psychological terms, are the respective advantages of the two methods? 2. Do the objective results coincide with your ejcperience in mem- orizing? If not, discuss the discrepancy. 3. If there was improvement with practice, try to explain on the basis of introspection the cause of the improvement. 4. Give directions for efficient memorizing. 5. How far do you think memory in general can be improved? 6. What are the limitations upon the conclusions to be drawn from a short experiment like this when applied to the general conduct of mental work? 7. Is there a contrast between the fast and the slow memorizers in their preference for the two methods? 8. Which method would be more likely to improve rapidly under intensive training? 9. Does this contrast between the methods of memorizing in the strict sense of the word apply to study in the broader sense? REFERENCES Ebbinghaus, H. Memory. Tr. by Ruger. Teachers College Pub- lications. Meumann, E. The Psychology of Learning. Tr. by Baird. Ap- pleton. Watt, H. J. Economy and Training of the Memory. CHAPTER III EXPERIMENTS WITH THE SCHOOL SUBJECTS Experiment No. 7 experimental psychological analysis of handwriting Problem : Analysis of the writing movement. As in the study of the learning process we began with a motor type of learning, so in the study of the psychol- ogy of the school subjects, we may begin with one which is predominantly motor in character. Such a subject is handwriting. From the point of view of the observer the writing process is a motor coordination. That is, it consists in a complex organized movement which is made up by the cooperation of a number of simpler movements. These simpler movements work together simultaneously in the production of the com- plex movement; and at each successive moment one particular combination of movements is followed by a different combination. In short, the simpler move- ments work together harmoniously, both simultane- ously and in succession. This experiment will consist in an objective analysis of the writing coordination into several of its simpler components. Material and method. The analysis of the move- ment into the contributory movement of the arm (in- EXPERIMENTS WITH SCHOOL SUBJECTS 73 eluding movements about the elbow and shoulder, the wrist, and the fingers, may be made by means of an adaptation of the apparatus used in Experiment No. 1. In place of the handle, used in the former experiment, there is a rubber band which is to be placed around the palm of the hand in front of the thumb and just back of the third joints of the fingers. The point of at- tachment with the rod which transmits the hand move- ment to the levers should be situated at the highest point of the band. The tracer record of the hand is reversed, but may be compared with the writing by turning it upside down. The experiment may be carried on as follows : First, place the two sheets of paper for the written and the tracer record with the edges parallel to the sides of the apparatus, and adjust the tracer so that it is comfort- able and follows the movements of the hand closely. Now take a record of several words, written in one's ordinary speed and manner of writing. To make com- parison easy, the same sentence may be written by all; for example, " A quick brown fox jumps over the lazy dog." Compare the tracer record and the writing so as to answer the following questions: — 1. To what component (element) of the total writ- ing-movement does the arm and hand contribute most? To what element do the finger movements contribute? 2. To what extent does the arm contribute in the formation of the letters? In this respect the indi- 74 EXPERIMENTAL EDUCATION vidual differences among the members of the class should be particularly noticed. 3. Distinguish the movement at the wrist from the movement at the elbow or shoulder, if such dis- tinction is shown by the record. Make another record, this time writing with as little finger movement as possible. a. To what extent were you able to exclude the finger movements? b. What modification, if any, was made in the appearance of the writing by the difference in the type of movement? Make a record with the tracer of a series of ten or more straight up-and-down strokes to resemble a series of saw teeth. Then make a series of dots on the paper on which the writing is done in the same relative po- sition as the upper and lower limits of the up-and- down strokes previously made, and make another record with the tracer of a series of up-and-down strokes, using the dots as limiting points. Compare the tracer records in the two cases and draw con- clusions. Treatment of results. The results of this experiment are not subject to exact quantitative treatment. The relative amount of hand movement used in the forma- tion of the letters may be estimated by comparing the height of the letters themselves with the height of the corresponding part of the tracer record. The same may be done for the " saw tooth " movement. In de- EXPERIMENTS WITH SCHOOL SUBJECTS 75 termining individual differences, the records may be put into classes roughly divided under such rubrics as little, medium, and much, — as for example in deter- mining individual differences in the amotmt of arm movement used. The correlation between the amovmt of finger movement and the quality of the writing may be worked out. • Results of the experiment. The more significant results of this experiment cannot be expressed in tabu- lar form. The report of such an experiment consists in the reproduction and interpretation of typical records. Three such records are shown in Figures 4 and 5. The following points should be observed and discussed in the report : — It is evident there is wide individual difference in the amount of arm movement in comparison to finger movement in the three cases whose records are before us. In the case in which the finger movement is least in amount there is still a certain amount, particularly in the earher words of a sentence. It is of importance to note that the additional amount of arm movement which can be introduced voluntarily is limited. Long- established habits are not quickly changed. The tracer record is shorter than the written line in all cases, but more so in some cases than in others. Why is this? It is not due to the failure of the pen to follow the movements of the hand attachment. To work out this mechanical problem the various parts of the hand and arm should be diagramed. The solution of the 76 EXPERIMENTAL EDUCATION problem throws light on some of the details of the movement. The slope of the line of writing and the slant of the / ^ 4 -^ a< u ft«*«-«is •t(»^"-K ^■^ y>^fj*J^ t ^d^ (P t>«^ (lr«'t™^*'£^^,5Lo« Flo. 4. TRACER RECORD FROM A WRITER WHO TJSES LITTLE (A) AND MUCH (B) ARM MOVEMENT a. Written in the writer's usual manner b. Written in an attempt to emphasize arm moyemeut individual letters is not always the same in the tracer record and the writing itself. Sometimes the slope of the tracer record varies from word to word. What EXPERIMENTS WITH SCHOOL SUBJECTS 77 adjustment of hand, arm, or fingers explain these vari- ations? The three records used for illustration are not enough to form the basis for conclusions regarding the Fia. 5. TRACEE RECORD FROM A WRITER WHO USES A MEDIUM AMOUNT OF ARM MOVEMENT a. Written in the writer's usual manner b. Written in an attempt to emphasize arm movement relation of the character of the coordination to the character of the written product, though they offer suggestions on this point. We have been taught to as- 78 EXPERIMENTAL EDUCATION sociate arm movement with a particular style of writ- ing which is taught in business colleges, and is being adopted in the schools. Such an association does not seem to exist in the specimens before us. This problem is of major importance, and should be studied more thoroughly on the basis of wider results from the whole class. The results of the experiment with the saw teeth are easy to interpret, and need not be illustrated here. An important part in any experiment in which ap- paratus plays a prominent r61e consists in testing the reliability of the apparatus. In the present instance, it is likely to seem to a good many writers who have learned the arm-movement style of writing that the apparatus does not faithfully represent their writing movement. The criticism may be raised that there is some play in the levers which makes the traced record less complete than the movement of the hand and arm. In an original experiment all such possible criticisms should be anticipated. Extensions of the experiment. One form of exten- sion of this experiment, then, may consist in determin- ing the degree to which the movement of the pen repro- duces the movement of the attachment of the band which goes round the hand. The test may be made roughly by holding the plate of the socket-joint firmly against a sloping pencil, and then tracing a figure with the pencil, taking care to keep the pencil at a constant angle to the paper and otherwise parallel to its position EXPERIMENTS WITH SCHOOL SUBJECTS 79 at the start. A more exact method is to construct a small three-legged stand, about three inches high and with rounded feet, to make it slide easily, and a vertical tube in the center large enough to hold a short pencil. The ball-and-socket attachment of the hand-band may then be fastened to the top of the stand directly above the pencil, in such a manner that the joint moves freely, and a record made by sliding the stand over the paper. If the tracer record does not correspond closely to the record made by the pencil the apparatus should be examined to see if any joints are loose. The two records should be very nearly identical. Another extension of this experiment, which may be made, is to conduct a practice experiment in the devel- opment of the arm movement, recording the progress by means of the tracer instrument. This will give some insight into the conditions of such a modification of a long-standing coordination, such as writing, and will give a basis for estimating the difficulty attendant upon the attempt to modify the handwriting habits of the pupils. The writing coSrdination of pupils themselves may be analyzed by means of the tracer. Comparisons may be made between the writing of pupils of different degrees of maturity, or between groups who have been taught by different methods. Another type of study of the writing coordination consists in a measurement of the speed and pressure- changes of the resultant total writing movement, in- 80 EXPERIMENTAL EDUCATION stead of the analysis of the component elements of the coordination. Such a study makes possible the deter- mination of such features as the rhythm of the writing movement. Rhythm is present when the successive strokes are made in something like equal periods of time. In the young child's writing long strokes are made in much longer time than short strokes, while in an adult's writing the duration of long and short strokes may be equal. The determination of such facts as these requires some form of apparatus by which the speed of movement may be measured. The writer's article in the Psychological Monographs, cited in the list of references, describes an elaborate form of appa- ratus which may be used for this purpose. A simpler form could be set up, particularly if the pressure were not recorded, to make less accurate measurements. On the other hand, the pressure alone could be re- corded by means of a relatively simple table, set upon a lever which was kept in position by a spring and records upon a kymograph drum. EXPEHIMENTS WITH SCHOOL SUBJECTS 81 QUESTIONS AND TOPICS FOR DISCUSSION 1. What conclusions may be drawn, if any, as to the proper func- tion of finger, hand, and arm movement? 2. Do the results throw any light on the amount of individual difference in this respect which it is desirable to allow? 3. What bearing does the fact regarding the function of arm movement have upon the most desirable position of the paper and slant of the writing. 4. Make any observations you can upon other acts of skill, and note individual differences in the coBrdination. 5. In what degree can the makeup of the coordination be con- trolled by the teacher or the learner? 6. Distinguish between fmm and execution in an act of skill. references' Freeman, F. N. "Some Issues in the Teaching of Handwriting"; in Elementary School Teacher (1911), vol. 12, pp. 1-7 and 53-59. Freeman, F. N. An Experimental Study of the Handwriting Move- ment, Psychological Monographs (1914), vol. 17, pp. 1-54. Judd, C. H. Genetic Psychology for Teachers, chap. VI. 82 EXPERIMENTAL EDUCATION Experiment No. 8 a test of handwriting Problem. The grading of handwriting, either for the purposes of teaching or of supervision, is very uncer- tain because of the lack of a means of accurately judg- ing its excellence. In making a judgment it is necessary to take into account all the characteristics which are essential to good writing. These include at least the speed with which the writing is produced and the qual- ity of the writing itself. The speed can easily be meas- ured by requiring the pupils to write, for a specified length of time, material which has been memorized. Quality is not so easy to measure, because of uncer- tainty as to what constitutes good quality and of the difficulty of finding means of measuring quality when it is defined. Thorndike, in constructing his measuring scale, used three characteristics for quality; namely, legibility, beauty, and character. Ayres used legibility alone. The two investigators agree in taking, as the basis of judgment, a general unanalyzed quality or group of qualities. The method of constructing their scales is different, but the method of applying them is the same; namely, to judge by the general impression which the writing makes. The scales help by giving standards by which these general impressions may be given numerical equivalents. The scale which is pre- sented for trial in this experiment differs from the two EXPERIMENTS WITH SCHOOL SUBJECTS 83 above mentioned in that one in using it does not at- tempt to compare the general impression made by the specimen to be judged with the impressions made by the specimens in the scale, but rather tries to analyze and rank the most important elements of the form of the writing which produce the impression of legibility, beauty, etc.^ The elementary characteristics of the writing which are assumed to constitute it legible, beautiful, etc., are four: (1) the degree of uniformity of the writing, (2) the quaUty of the line, (3) the correctness of letter forma- tion, and (4) the spacing or composition. It is neces- sary to explain what is meant by these categories more fully. Uniformity is applied particularly to two character- istics, slant and alinement, the latter including the height of the letters. It cannot be said that any par- ticular slant is the best, but it is clear that the slant of any person's writing should be uniform. Variability in slant may be of various sorts. In the first place, it may be due to a general lack of coordination, so that the variability follows no rule. In the second place, it may consist in an increased slant toward the end of the line, which is due to the lack of corrective adjustment. Finally, it may be due to occasional readjustments in * The Thomdike Scale can be obtained from the Bureau of Pub- lications of Teachers College, Columbia University, New York City; The Ayres Scale from the Russell Sage Foundation, Division of Edu- cation, New York City; and the Analytical Scale from Houghton MifiBin Company. 84 EXPERIMENTAL EDUCATION the position of the hand or paper, the slant remaining fairly uniform between these readjustments. Similarly, the letters should come to a common base line, and the same kinds of letters should be of the same height. We may confine our attention, for convenience, to the one- space letters. It is not easy to grade uniformity of slant and of alinement together, so these two categories have been separated in the charts. All the characterigjtics so far described might appear in a high degree of excellence, and yet the writing might be seriously 'at fault if the line itself were not clear and firm. The line should have evenness in width, indicating evenness of pressure upon the pen, or at least changes in width should not be great, abrupt, nor irregular. A moderate degree of even shading should not be penalized. The direction of the line also should not change in an uneven fashion, producing a wavy or jagged appearance and indicating lack of freedom, ease, and control of movement. The third category, therefore, is quality of line. The fourth characteristic refers to the arrangement of the lines within the letter itself. In general the point to be judged is the degree of conformity of the letter formation to the standard letters. In applying this standard the judge must avoid using the peculiarities of any particular style of alphabet as a guide, and must try to distinguish the requirements on which all English alphabets would agree. For example, all would agree that the " o " should be closed, that the last EXPERIMENTS WITH SCHOOL SUBJECTS 85 stroke of the " a " should come to the base line, that the " t " should not have a loop, etc. Perhaps the best way to judge this characteristic is to look through the writing and try to estimate rapidly the degree to which the letters are well formed and clear. The fifth category, spacing, is somewhat more diffi- cult to analyze and rank. The aim from this point of view may be said to be — first, to so place the letters and words on the page that the words themselves pos- sess unity. The letters should be far enough apart to make them easily distinguishable, but should be close enough together to give the word a compact appear- ance. In the second place, the words should be related to each other much as the letters are related in the word. There should be enough space between the words on a line, and between the lines, to render each word distinct, but no more. The specimens on the chart illustrate the application of these principles. Material and method. In order to aid in discrimi- nating these different characteristics, and awarding to specimens grades according to their standing in each quality, a chart has been prepared. This chart is composed of five series of writing specimens, arranged in an ascending scale of excellence, each series repre- senting one of the qualities which have been described. Thus the first series represents different degrees of uniformity of slant; the second series, degrees of uni- formity of alinement and size; etc. Three degrees of excellence are distinguished in each chart. In all the 86 EXPERIMENTAL EDUCATION categories these are given scores of 1, 3, and 5 respec- tively. The intermediate scores 2 and 4 may be used. The scores for letter formation may be doubled, on the assumption that this characteristic is more im- portant than the others. It is probably better to use the same scoring as in the others in making the judg- ments, and then double them afterwards. The specimen to be judged is graded according to each category separately, and given the rank of the specimen in the chart with which it most nearly cor- responds in each case. The total rank is calculated by summing the five individual ranks. Thus, if letter formation is given double value, the lowest possible rank is 6, and the highest possible rank is 30 (5, plus 5, plus 5, plus 10, plus 5), and the range is 24. Several precautions are to be observed in making the judgments. The value of the method rests upon the fact that different features of the writing are singled out one at a time, and are graded by being given a rank in one of only three steps. The differences between the steps are marked, and the ease of placing a specimen should be correspondingly easy. This method implies, however, that (1) The attention is fixed on only one characteristic at a time. (2) The judgment on one point be not allowed to influence the judgment on the other points. (3) The same fault be counted only once. (4) General impression be disregarded. EXPERIMENTS WITH SCHOOL SUBJECTS 87 These four rules, in fact, express different aspects of the same precaution, but it is worth the emphasis thus given. The experiment consists in judging, by means of the method which has been described, one hundred spe- cimens of writing of different styles and from writers of different ages. (In order to give some preliminary practice, ten additional specimens should be provided which are to be graded first.) No marks are to be made on the specimens themselves. Each one is given a serial number, and a record is to be kept of the indi- vidual ranks as well as the final rank assigned to each specimen. The individual reports should contain a table con- taining these data, together with the average rank as- signed to the group of papers. Each individual should also put the ranks which he has assigned into a chart to show their distribution. (For example of a distribu- tion chart, see the article by Starch and Elliott, in the School Review for April, 1913, p. 256.) In this case the units on the base line represent the possible ranks from 6 to 30. Above each rank should be written the num- bers of all the papers which are assigned that rank. Results of the experiment. The results obtained by the use of the analytic scale with untrained judges have not as yet warranted a positive statement as to the possibility of obtaining highly consistent results in the grading of writing by different graders. We may first examine the record of the scores given by one group of 88 EXPERIMENTAL EDUCATION ten graders to 50 specimens of children's writing, and then consider their significance. In Table VII are given the scores given by the ten graders to five of the papers, chosen so as to represent different degrees of excellence. TABLE VII. SCORES OF TEN INDIVIDUALS GRADING FIVE WRITING SPECIMENS Grader Specimen 1 2 3 4 5 A 8 7 8 6 7 10 9 9 12 8 14 13 10 10 19 18 15 17 14 18 14 15 16 22 21 17 17 14 18 20 18 14 20 12 22 28 24 21 20 24 28 B 23 C 28 D 28 E 28 P 30 G 29 H 29 I 29 J 26 Average 8.4 14.9 17.4 20.3 27.8 The examination of this table makes it clear that there is a good deal of variation in the grades which different untrained individuals give to the same pa- pers. In specimen 1 the range is from 6, the lowest possible rank, to 12. In specimen 4 the range is from 12 to 28. EXPERIMENTS WITH SCHOOL SUBJECTS 89 TABLE VIII. AVERAGE OF THE GRADES GIVEN BY TEN GRADERS TO THE SAME FIFTY PAPERS, USING THE ANALYTIC AND THE AYRES SCALES Graders Averages by analytic scale Variations from gen- eral average Averages by Ayres scale Variations from gener- al average 1 18.2 13.7 16.34 15.86 17.08 19.05 19.86 19.8 19.58 19.64 .3 4.2 1.6 2.0 .8 1.2 2.0 1.9 1.7 1.7 59.8 52.4 48.9 47.1 58.0 63.1 60.1 56.8 51.0 70.7 3 2 4 4 3 7 9 4 9 7 6 1 2 6 6 3 7 3 3 8 9 5 8 10 13 9 Average 17.9 56.8 Mean Variation . . •• 1.74 •• 5.6 The examination of the variation among the average of the grades given by the different graders to the whole set of fifty papers may be more encouraging. These averages are shown in the second column of Table VIII. The mean variation of these averages is 1.74 steps on the scale, the extreme range of which is 24 points. This variation is in both directions from the average, so that about half of the graders are 3.5 points or more apart in the average rank they assign to a group of fifty papers. For instance, in this group the third grader from the bottom assigns an average grade of 16.34, and the second from the top 19.8. This dif- 90 EXPERIMENTAL EDUCATION ference of 3.5 points is about equivalent to the average progress made in two school years. In the case of an- other group of fifteen graders the variation was con- siderably less, being 1.49. A study has already been made of the variability of graders using the Ayres scale. See the article by Man- uel, cited in the list of references. In the present exper- iment a similar test was made and the final results are shown in the last two columns of Table VIII. The question which immediately arises is. Which of the two variabilities, 1.74 or 5.6, is the greater? We cannot determine merely on the basis of the amoimts taken at their face value, since they are based on en- tirely different imits. Various means of equating such coefficients of variability have been used or suggested. See the articles by Starch,* Weiss,^ Pintner,' and Kelley.^ The most reliable method is the one which takes, as the measure of the units which are used, the actual range among specimens which are graded in terms of these units. If the same specimens are graded by two scales, and are distributed according to the scores ' Starch, D. "The Measurement of Handwriting"; in /ourno/ of Educational Psychology (1913), vol. 4, pp. 445-64. ' Weiss, A. P. "A Modified Slide Rule and Index Method in Indi- vidual Measurements"; in Journal of Educational Psychology (1914), vol. 7, pp. 190-225. ' Pintner, R. " A comparison of the Ayres and Thomdike Hand- writing Scales"; in Journal of Educational Psychology (1914), vol. 5, pp. 525-36. * Kelley, T. L. "Comparable Measures"; in Journal of Educa- tional Psychology (1914), vol. 5, pp. 589-95. EXPERIMENTS WITH SCHOOL SUBJECTS 91 which are given to them, we may assume that the range of one distribution roughly represents the same range in merit that is represented in the other. Since extreme cases are Hkely to be somewhat unrehable, we may take the extent of the middle half (approximately) of each distribution as representing equal ranges in merit. Thus, in the present case, the score of each of the fifty papers on the Analytic and Ayres Scales (based on the average of the rankings of the ten grad- ers) is as follows: — COMPARISON OF SCORES, USING Analytic Ayres scale scale 8.4 24.0 10.8 31.5 11.2 37.5 12.8 39.5 13.4 41.0 13.4 42.0 13.7 42.5 14.0 43.5 14.1 44.5 14.4 45.0 14.8 45.5 14.9 46.5 14.9 48.5 15.0 49.0 15.0 49.0 16.0 50.0 16.1 51.0 16.2 52.0 16.4 52.5 16.6 52.5 16.6 53.0 16.6 53.0 16.8 53.5 17. 1 54.5 17.1 55.0 Analytic Ayres scale scale 17.4 56.5 17.5 58.5 17.7 58.5 17.8 59.0 18.0 59.0 18.2 60.0 18.3 60.0 18.3 61.0 18.6 61.0 19.5 61.0 19.6 62.5 20.3 64.0 20.3 65.5 21.6 65.5 21.9 67.5 22.5 68.5 22.9 71.0 23.4 73.5 23.7 73.5 23.9 75.0 24.1 75.5 25.0 77.0 26.1 79.0 27.5 83.0 27.8 88.0 92 EXPERIMENTAL EDUCATION The scores are arranged in ascending order in each case, and the scores which correspond in position in the two series do not necessarily represent the same paper. The assumption in this particular case is that the range 14.9 to 20.3 is approximately equivalent to the range 48.5 to 65.5 or, in other words, that 20.3-14.9, or 5.4 units on the analytic scale, are approximately equivalent to 65.5-48.5, or 17 units on the Ayres scale. One unit on the analytic scale is then equivalent to 3.15 units on the Ayres scale. We are now in a position to equate the two coeflB- cients of variability, 1.74 and 5.6, by multiplying 1.74 by 3.15, and thus turning it into equivalent units. The variability turns out to be 5.48 by the analytic scale, as compared with 5.6 by the Ayres scale. Neither of these is satisfactory if we are to compare ratings made by different investigators, unless we have each set of papers rated by a group of judges. Prehm- inary tests have indicated, however, that a much more reliable set of judgments may be obtained by the ana- lytic method when the judges are trained for their task. This requires ample discussion by a group of graders upon the first few papers they grade. After a careful study and comparison of their grades upon twenty-five papers, the variation becomes inconsider- able. Practice, with conferences, also greatly reduces the variability in the use of the other scales. (See Gray's article referred to below.) EXPERIMENTS WITH SCHOOL SUBJECTS 93 Extension of the experiment. This is one type of extension of this experiment that could profitably be made. One might, if successful, go still further, and determine whether judges might be trained by using a set of printed instructions accompanied by a set of papers and their standard ratings. Further extensions of this experiment may be made by using some of the scales which have been developed in other subjects, and testing their reliability. The list of references below indicates directions in which fur- ther study might be made. Finally, the attempt may be made to work out and standardize an analytic scale for the measurement of attainment in some other sub- ject of the curriculum. QUESTIONS AND TOPICS FOR DISCUSSION 1. Discuss fully the adequacy or inadequacy of this method of testing handwriting. a. Were any categories less satisfactory than others? 6. Is the list of important categories complete? c. Are any particular specimens unsatisfactory? d. Is the use of the method unduly cumbersome? e. Compare the method with those of Thorndike or Ayres if you have used them. 2. How might such procedure lead to a better appreciation of the qualities of good writing? 3. How might such a method help the teacher in teaching? 4. How reliable is the method for comparing different grades or schools? (The general report should include an especially full discussion of this question.) 5. Add any suggestions which occur to you. 6. Does an analytical scale require any more training for its satis- factory use than one which requires judgment from general impression? 7. Is there any other method than that used by Gray (Texas 94 EXPERIMENTAL EDUCATION bulletin) of determining the relative importance of the vari- ous categories? What is it? 8. Look up, if you can, the method of grading some agricultural product and give a sample score card. REFERENCES Ayres, L. P. A Scale for Measuring the Quality of Bandwriting of School Children. Russell Sage Foundation, Division of Education, New York City. Freeman, F. N. "Problems and Methods of Investigation in Handwriting"; in Journal of Educational Psychology (1912). Freeman, F. N. The Teaching of Handviriting. Houghton MifiBin Company (1914). Freeman, F.N. "An Analytical Scale for Judging Handwriting"; in Elementary School Journal (1915), vol. 15, pp. 432 -41. Gray, C. T. "The Training of Judgment in the Use of the Ayres Scale for Handwriting"; in Journal of Educational Psychology (1915), vol. 6, pp. 85-95. Gray, C. T. "A Score Card for the Measurement of Handwriting,'' Bulletin of the University of Texas (1915), no. 37. Manuel, H. T. "The Use of an Objective Scale for Grading Handwriting"; in Elementary School Journal (1915), vol. 15, pp. 269-78. Thorndike, E. L. "Handwriting"; in Teachers College Record (March, 1910). EXPERIMENTS WITH SCHOOL SUBJECTS 95 Experiment No. 9 observation of eye movements in beading Problem. In reading, as in writing, we shall begin with the study of the motor coordination, as the proc- ess of reading depends upon the formation of motor habits, just as does writing. Material and method. For the accurate study of eye movements it is necessary to re- sort to some me- chanical means of recording them, such as the me- chanical recording device described by Huey, or some pho- tographic method. The outstanding characteristics of the movement may be observed, however, without the use of these refined methods. In fact, Erdmann and Dodge, who were the pioneers in the recent study of the eye movements in reading, merely observed the eye through a telescope. In- stead of using a telescope we may observe the eye Flo. 6. MIRROR rOR OBSERVINa EYE MOVEMENTS 96 EXPERIMENTAL EDUCATION movements in a mirror from behind the subject, as is shown in Figure 6, and thus facihtate observa- tion by getting close to the eye which is to be ob- served. The mirror is set in an adjustable frame so that it may be close to the subject's eye, and at such a height that the top of the mirror is on the level with the lower edge of the pupil of the eye when the head is held erect. With this arrangement the subject can look over the mirror and read material which is held at a level with the eye, and at the same time the experimenter can see enough of the eye to observe its movements. The sub- ject should look toward the window, or other source of light, and the text which is to be read should be placed about on the level of the eye. It requires a little prac- tice to catch the small and rapid eye movements, and the method is at best a means to an approximate count of the movements which are made in reading each line of print. The subject should begin at a known place on the page, and look from the left to the right margin before beginning to read, for the purpose of giving a warning signal to the experimenter. The speed of reading should be measured by means of a stop-watch in the hand of the subject. The experiment ter should, without taking his eyes from the eye of the subject, make note with pencil and paper of the num- ber of fixations made in reading each line. A frequent som-ce of error is to count the eye movements, instead of the fixations. When this is done, either the return EXPERIMENTS WITH SCHOOL SUBJECTS 97 sweep must be counted or 1 must be added to the count for each Une. The subject should read silently and continuously. The number of pauses to a line, and the average number of words to a pause, should be calculated; first for the same sort of subject-matter, but for different lengths of line and different sizes of print. For this purpose there are first provided two texts, the subject- matter of which is printed in 11 -point type; the first, No. 1, having a length of line of 24 ems, and the other. No. 2, of 12 ems. In order to study the effect of the size of type on the eye movements, a third text, having a length of line of 24 ems and printed in 7- point type, is provided for comparison with the first specimen. In order to test the effect of the character of the subject-matter upon the eye movements, a fourth specimen, No. 4, is provided, having the same type and length of line as specimen 1, but having nar- rative as subject-matter. Finally, to test the speed of reading upon eye movements, a fifth specimen. No. 5, similar to specimen 1, in type, length of line, and sub- ject-matter is provided, which is to be read as rapidly as possible. All of these texts are reproduced in the Appendix, which see. Treatment of results. Each individual should cal- culate the average number of pauses per line and per unit length of line; the average number of words per line, and the average number of words perceived dur- ing one reading pause for each of the five texts. The 98 EXPERIMENTAL EDUCATION TABLE IX. RESULTS FROM THE STUDY OF READING PAUSES TEXT I Individual No. pauses per line Ave. no. pauses per em Ave. no. words per line Ave. no. words per pause Ave. no. pauses per second Ave. no. words per second 1 4.88 2.89 4.64 2.64 3.6 5.0 2.37 3.8 5.1 5.7 3.97 7.07 4.6 3.4 .20 .12 .19 .11 .14 .20 .10 .15 .21 .24 .16 .29 .19 .14 10.8 10.0 11.2 11.3 11.0 11.3 10.0 11.8 11.3 11.3 11.0 11.2 11.2 11.0 2.2 3.62 2.41 4.17 3.4 2.3 4.2 3.4 2.2 2.0 2.77 2.43 2.48 3.05 2.1 1.61 1.08 1.84 1.17 2.78 2.0 1.82 2.0 1.60 1.82 2.87 2.0 2.35 4 81 2 6.31 3 2.45 4 7.68 5 3.67 6 6.28 7 8.41 8 6.91 9 4.4 10 3.0 11 6.06 12 4.64 13 4.32 14 7.19 Average 4.25 .17 11.0 2.81 1.92 5.21 EXPERIMENTS WITH SCHOOL SUBJECTS 99 TABLE IX (continued) TEXT II Individual No. pauses per line Ave. no. pauses per em Ave. no. words line Ave. no. words per pause Ave. no. pauses per second Ave. no. words per second 1 2.94 1.09 3.9 1.17 2.0 2.7 .8 2.4 3.0 3.1 2.05 3.37 3.0 1.94 .24 .09 .32 .09 .16 .22 .07 .20 .25 .26 .17 .28 .25 .16 5.08 5.0 5.1 4.91 5.0 4.8 5.0 5.0 4.8 5.0 5.0 5.1 5.08 4.91 1.7 4.61 1.30 4.19 2.50 1.8 6.3 2.1 1.6 1.6 2.43 1.48 1.69 2.53 2.7 1.27 1.82 1.95 1.27 3.15 1.3 2.62 2.7 1.6 2.18 2.12 2.86 3.22 4.7 2 5.83 3 2.38 4 8.19 5 3.18 6 5.6 7 8.33 8 5.44 9 10 4.3 2.5 11 5.3 12 3.88 13 4.94 14 8.15 Average 2.39 .20 4.98 2.56 2.23 5.05 100 EXPERIMENTAL EDUCATION TABLE IX (continue^ TEXT m Indimdual .Yo. pauses per line Ate. no. pauses per em .-Ire. no. uords per line Ate. no. trords per pause Are. no. pauses second Ate. no. words per second 1 6.34 4.49 7.09 3.11 4.5 5.83 3.41 3.72 6.4 6.4 5.58 8.04 8.37 4.63 .26 .19 .29 .13 .20 .24 .14 .15 .27 .27 .23 .33 .34 .19 15.5 16.7 14.9 15.3 15.0 15.0 16.7 14.7 15.0 14.9 15.0 14.9 15.4 15.3 2.4 3.72 2.6 4.56 3.33 2.57 4.9 4.0 2.30 1.85 2.68 1.85 1.84 3.29 2.6 1.76 1.32 1.8 1.11 3.12 2.2 1.3 1.9 1.11 2.14 2.3 2.5 2.4 6 4 2 6 53 3 2 78 4 8 22 5 3. 7 6 S 06 7 10 9 8 5 13 9 4 3 10 2.7 11 12 4 3 13 4 61 14 7 92 Average 5.56 .23 15.3 3.03 1.97 5.81 EXPERIMENTS WITH SCHOOL SUBJECTS 101 TABLE IX (continued) TEXT IV IndimdmU No. pauses per line Ave. no. pauses per em Ave. no. icords per line Ave. no. toords per pause Ave. no. pavses per second Ave. no. words per second 1 4.91 3.12 5.06 1.88 3.5 4.33 2.48 3.03 5.3 5.1 3.88 6.87 5.28 3.11 .20 .13 .21 .08 .14 .18 .10 .12 .22 .21 .16 .28 .22 .13 12.4 12.0 12.8 12.3 12.0 12.3 12.0 12.5 12.3 12.5 12.0 12.8 12.4 12.4 2.5 3.85 2.53 6.56 3.42 2.84 4.8 4.1 2.3 2.4 3.09 1.86 2.16 3.97 2.9 1.75 1.6 1.53 1.09 3.03 2.35 i.l 2.4 1.9 2.61 2.67 2.66 2.57 7 25 2 6 74 3 4 07 4 10 07 5 3 71 6 8 61 7 11 4 8 8 7 9 5 5 10 4.6 11 8 01 12 13 4.97 5 98 14 12.1 Average 4.13 .17 12.3 3.31 2.23 7.27 102 EXPERIMENTAL EDUCATION TABLE IX (continued) TEXT V X Individual No. pauses per line Ave. no. pauses per em Ave. no. words per line Ave. no. words per pause Ave. no. pauses second Ave. no. words per second 1 4.31 3.4 5.44 2.06 3.67 3.46 2.2 3.09 5.4 4.2 3.82 5.51 5.0 3.6 .18 .13 .22 .08 .13 .14 .09 .13 .22 .18 .16 .23 .20 .15 10.7 10.3 10.7 10.6 11.0 11.3 10.3 10.8 11.3 10.6 11.0 10.7 10.6 10.6 2.4 3.29 1.96 5.16 3.59 3.26 4.7 3.4 2.1 2.5 2.87 1.81 2.13 2.89 2.6 1.67 1.42 2.0 1.6 2.96 2.2 2.16 3.6 2.2 3.52 2.57 2.62 2.38 6.2 2 5.47 3 2.79 4 10.3 6 6.38 6 9.64 7 10.3 8 7.4 9 7.6 10 5.7 11 10.13 12 13 4.66 5.05 14 6.82 Average 3.93 .16 10.7 3.0 2.38 6.96 EXPERIMENTS WITH SCHOOL SUBJECTS 103 average number of pauses made per second and the number of words read per second should also be cal- culated. Note should also be made of the ease or diffi- culty, comfort or discomfort, and attractiveness or im- attractiveness of the various texts. The general report should include a table in which these facts are brought together in such a way that individual differences and averages shall appear. These results should then be interpreted. Results of the experiment. A summary of the nu- merical results from one section of fourteen individuals in Experiment No. 9 is presented in Table IX. A word should be said regarding the degree of accuracy of these results to enable one to discriminate between those conclusions which are justified and those which are not. In the first place, the count of the ntunber of pauses per line, which is the basis of all the formulation of results, is probably in all cases, and clearly in some, inaccurate. Note, for example, the record of eight tenths of a pause per line for Individual 7, in Series II. Though the lines in this series were short, it is contrary to all the evidence which has been obtained by an accm-ate method of recording to suppose that a person can read more than a line per pause. This, and other low records, is probably due to two errors: first, the failure to catch some of the eye movements; and sec- ond, counting the movements within a line instead of the pauses, resulting in two pauses being counted as one. The results then are, in general, lower than they 104 EXPERIMENTAL EDUCATION should be in pauses per line, but we cannot tell how much. We cannot from these results, then, draw any definite conclusions regarding the absolute nimiber of pauses per line or any of the facts which are derived from them. A further diflBculty is due to the variability in accu- racy among the different observers. It is certain that some fell into the error which has been mentioned of counting movements instead of pauses, and there are undoubtedly differences in accuracy besides this. This makes it impossible to make any certain statements about individual differences on the basis on these data, except in reference to the relation of the different series to one another. Furthermore, there are some inaccu- racies in the individual calculations themselves, as is shown by the difference in the calculation of the num- ber of words per line, and by the failure of some mem- bers to check which should. For instance, the number of words per second should be the product of the num- ber of words per pause and the number of pauses per second, but it is not so in all cases. This is a type of inaccuracy which ought not to occur. Notwithstanding these errors, there is a fair degree of reliability in the comparison of the results obtained with the successive texts; since some of the errors are fairly constant, and the others, in the average results of a group, counterbalance one another. One other possible error which would affect the comparison of earlier series with later ones is an increase in skill in the EXPERIMENTS WITH SCHOOL SUBJECTS 105 observers, enabling them to detect more fixations as they become more practiced. This would result in relatively too many fixations or pauses in the later series. This does not seem to be true to an appreciable extent, however, since the pauses per em, or unit length of line, are fewer in the last series than in the first. We may hold that they should be still fewer than they are in comparison, but it seems likely that this source of error is not serious. We may then proceed to compare the data regarding eye behavior for the group under the conditions of the five series. Take first the final product of the various factors, the rapidity of reading as expressed in terms of words read per second. Of the four texts which were read without any especial attention to speed, the fiction is read most rapidly by each individual, considerably so by most. On the average it is read more rapidly than the scientific text. No. 5, which is read rapidly inten- tionally, though there is not much difference in speed between reading fiction and intentionally rapid reading of scientific material. Of the three selections which differ in length of line or size of type, but are alike in subject-matter, that which is printed in small type is read most rapidly, on the average, and by eight out of the fourteen individuals. Between Texts I and II there is not much difference, the average being slightly in favor of the longer line. What now are the factors in these differences in the rate of reading? Compare first fiction with the other 106 EXPERIMENTAL EDUCATION materials. When we examine the data in detail we discover that the more rapid reading of this text is partly due to a fact which decreases its significance, namely to the fact that the words are shorter, as is in- dicated by the larger number in a line as compared with the other texts of the same type and length of line. We should then expect more words to be read per second. The number of words per line also affects the number of words read per pause. The number of pauses per line is practically the same as in reading the first text, and hence if the words were of the same length the number of words read per pause would be the same. There is no superiority manifested, then, in the span of attention in reading fiction material. There is a slight superiority, however, in the frequency of the pauses themselves, as indicated by greater number of pauses per second. If this should be confirmed by more careful investigation it would indicate that ease of apprehension serves to lessen the diu-ation of eye pauses rather than to increase their scope. If we examine the data for Text V, for intentionally rapid reading, we see that the gain was made by both an increase in the scope of apprehension at each pause (decrease in the number of pauses per line or per em) and a decrease in the duration of the pauses (increase in the number of pauses per second) . If we assume that the increase in the rapidity of reading of this text is due to a greater output of mental energy, it appears that such an increased output expresses itself both in increased scope and in decreased duration. EXPERIMENTS WITH SCHOOL SUBJECTS 107 It only remains to notice the effect of length of line and type. The shorter line appears to result in a greater number of pauses per imit distance, but this is partly counterbalanced by somewhat lessened dura- tion of the pauses. The greater number of pauses seems to be easily explicable by the fact that at the end of the line there is likely to be waste, due to the fact that the amount left over from the next to the last pause is not enough to supply the number of words for a normal fixation, but is enough to necessitate an addi- tional one. If we multiply the number of ends of lines we obviously multiply the number of chances for this condition. These results and this explanation should be compared with Dearborn's discussion of the length of line and its effect upon perception in reading. The finer type seems to allow somewhat greater scope of fixation, but the number of fixations per sec- ond does not increase. Extension of the experiment. Two ways readily suggest themselves in which this experiment may be amplified or extended. In the first place the results may be gained by a more accurate method. This neces- sitates the use of some sort of accurate recording de- vice. The two general methods which have been used are the attachment of a lever to the eye by placing a very light cup on the cornea and connecting it with the lever by a fine thread. This method was used by Huey. The other refinement of method uses photography. Dodge and Dearborn photographed the bright spot on 108 EXPERIMENTAL EDUCATION a falling plate. The kinetoscopic method, used by Judd in studing other eye movements, is also applic- able in the study of eye movements in reading. The photographic method has entirely superseded the mechanical lever method on account of its safety, and the fact that it interferes less with the natural eye movements. Any of these refined methods require somewhat elaborate apparatus and the mastery of a moderate degree of mechanical technique. This is an illustration of the fact that, in the laboratory investigation of psy- chological and educational problems, the ability to de- vise and manipulate apparatus of a suitable sort and of a sufficient degree of delicacy is often essential. If a student is interested in specialization in educational psychology, and the laboratory offers the requisite facilities, this experiment could well be extended by taking some photographs of eye movements in reading. The other mode of extension of this experiment is to carry further the method of observation. Other texts may be used to compare a variety of kinds of subject- matter. The method may be refined by having one observer go through a period of training in observation, and then observe the eye movements of a nvunber of the other students. This will give some check upon the variability among the various observers. EXPERIMENTS WITH SCHOOL SUBJECTS 109 QUESTIONS AND TOPICS FOR DISCUSSION 1. Describe the process of perception in reading so far as the char- acter of the eye movements gives evidence concerning it. i. What evidence is there from the experiment as to : a. The influence of external conditions on the eye movements? 6. The influence of the apprehension of meaning on the eye movements? 3. How far does it seem likely that development of eye-movement habits is an important factor in learning to read? 4. Consider the question of most favorable size of type, length of line, and arrangement of line in printing. REFERENCES Dearborn, W. F. Psychology of Reading. Huey, E. B. Psychology and Pedagogy of Reading. Judd, C. H., McAllister, C. N., and Steele, W. M. "Introduction to a Study of Eye Movements by Means of Kinetoscope Photo- graphs"; in Psycholgical Review, Monograph Supplement (1905), vol. 7, pp. 1-16. 110 EXPERIMENTAL EDUCATION Experiment No. 10 study of the perceptual process in reading by the tachistoscopic (or brief exposure) METHOD Problem. In reading, as in other products of per- ceptual development, expertness is reached by the recognition of increasingly complex objects. The char- acter of the eye movements which were studied in the previous experiment indicates that in adult reading numbers of letters and even words are grouped or fused in the recognition which occupies a reading pause. In this experiment this fusion is to be studied by compar- ing various cases in which it exists in dififerent degrees. Material and Method. Since a reading pause con- sists in a very brief fixation of the eye, we may artifi- cially produce reading pauses by exposing reading matter for a brief interval of time, and thus study the successive reading acts in isolation. In particular we are able to determine how much of different kinds of material may be recognized in each of several succes- sive exposures, and thus how the recognition may be built up. This method may be used to study the differ- ences in fusion in the recognition of different kinds of material, on the assumption that the greater the fusion between the elements of the perception the greater will be the amount which can be recognized in a given ex- posure, or the fewer will be the number of exposures necessary to recognize the whole. EXPERIMENTS WITH SCHOOL SUBJECTS IH The apparatus is one of a number which have been devised to expose objects to view for a brief time. Be- cause of the fact that the exposure is made by a falling screen, the apparatus is called a fall tachistoscope. An opening in the screen exposes the stimulus card for a length of time depending on the size of the open- ing and the height from which the screen drops. Any of the commonly used tachistoscopes will serve the purpose. A small instrument in which the screen is operated by a spring instead of by gravity, devised by Professor Dearborn of Harvard, is especially to be recommended because of its compactness and slight noise. In making each exposure the experimenter should make sure that the subject is in a favorable position and ready to pay attention. He shoidd then give a warning signal, followed in about one second by the exposure. He should be careful to keep the interval between the warning signal and the exposure as con- stant as possible. The subject should write down what he was able to recognize after each exposure, and should not give his reply aloud to the experimenter. The experimenter should not look at the face of the cards. The exposures of each card should continue until the subject is satisfied, but he should not be told whether or not his recognition is correct. The next card may then be shown. The subject should make introspective notes. We shall first compare the amount of fusion which 112 EXPERIMENTAL EDUCATION exists in the recognition of groups of digits and groups of the same number of letters in short words. The Hsts are given in the Appendix, which see. The second comparison is between groups of non- sense syllables and groups of short words arranged in a sentence, and containing approximately the same number of letters. The third comparison is between the recognition of familiar and unfamiliar words. The fourth group of words is for the purpose of studying the " proof reader's fallacy " — that is, the neglect of errors in words. The existence of this neglect is an evidence of fusion or recognition of the group as a whole, rather than of each unit for itself. The subject should not know which words contain errors, nor should he be required to detect them or make special effort to recognize details. Treatment of results. The results may be made comparable by dividing in each case the number of objects (letters or digits) exposed by the number of exposures necessary to recognize them, to give what we may for convenience call the average number of objects recognized per exposure. The larger this num- ber the greater we may assume the fusion among the objects to have been. The individual reports should include a calculation of this number of objects recog- nized per exposure for the different classes of stimuli, with some interpretation of the result. The general report should generalize the individual results and determine individual differences. EXPERIMENTS WITH SCHOOL SUBJECTS 113 TABLE X. THE AVERAGE NUMBER OP LETTERS OR DIGITS RECOGNIZED PER EXPOSURE (Based on the total number of letters or digits exposed, divided by the number of exposures necessary to recognize them) Serieg Individ- ual 1 1 1 as i-S n i II 11 1-1 ^ A B C D E F G H I J K L M N P Q R 1.5 1.2 .9 1.5 1.5 1.2 1.5 2.6 2.7 1.9 2.4 1.5 1.9 1.8 2.0 2.0 1.7 2.0 4.3 5.0 5.2 6.0 4.5 4.0 6.0 5.5 4.7 6.0 5.0 5.1 5.0 5.8 6.0 6.0 5.7 6.0 2.0 2.1 1.1 2.1 2.9 2.0 2.9 2.4 1.2 2.3 2.7 2.0 3.6 3.8 3.5 3.7 3.0 3.1 4.3 4.4 4.3 5.9 6.3 6.5 7.6 7.2 7.2 8.3 6.9 7.6 10.8 10.1 8.2 13.0 15.0 9.6 1.2 3.4 .7 1.8 1.2 1.8 2.6 1.6 6.7 3.9 2.6 8.9 3.4 3.3 6.5 4.3 8.7 13.0 5.6 7.0 12.0 7.0 10.0 14.0 10.5 12.0 14.0 14.0 14.0 14.0 14.0 14.0 14.0 14.0 13.0 14.0 3.0 3.6 4.5 5.5 5.4 4.3 5.0 5.5 6.3 5.4 8.5 8.5 7.7 8.4 8.5 8.0 7.3 7.3 3.1 3.8 4.1 4.3 4.5 4.8 5.2 5.3 5.8 5.9 6.0 6.1 6.6 6.7 7.0 7.3 7.8 7.9 Average 1.7 5.3 2.6 8.0 3.9 12.1 6.3 5.7 Results of the experiment. The summary of the results of a group of eighteen individuals is given in Table X. The main facts are evident without the ne- 114 EXPEEIMENTAL EDUCATION cessity of a minute analysis. A comparison of both the individual and the average results of the various series or kinds of stimuli makes clear the wide difference in the range of objects which can be recognized at a glance, or which can be apprehended simultaneously, according to the degree to which the elements are or- ganized. These results do not in every case indicate the limit of simultaneous apprehension for a particular kind of stimulus, since the groups (words or sentences) were not in all cases large enough to reach the limit of ability of some of the individuals. It must be remem- bered also that most of the individuals were unprac- ticed in tachistoscopic work. Wide individual differences are also brought out by the results, though these are to be somewhat dis- counted by differences in procedure of the various ex- perimenters. In general, the individual differences are similar in the various series, indicating that a similar factor is present in the recognition of the various kinds of objects. These differences may be subjected to further analysis by turning the individual scores into percentages of the average, and noting the percentage differences between the extremes in each series. The correlations between the standing in the various series may also be worked out. One of the graphic methods, to be described in connection with Experiment No. 16, would be useful for this purpose. "^ There is no evidence ^ See the chapter in Whipple for references to Messmer, Meu- mann, and others who have made a distinction between objective and subjective observers. EXPERIMENTS WITH SCHOOL SUBJECTS 115 in these results to support the view that there are clearly separated types, however we may interpret the individual differences which appear. Extension of the experiment. It is a profitable exer- cise to identify the individual subjects in the various experiments, and determine how far individual differ- ences in certain of those which are similar show corre- lation. The results of this experiment may thus be correlated with the results of Experiment No. 12, " Apprehension of Number." The experiment may be extended by using still other kinds of objects, such as those of the spot pattern test (see Whipple for description). Children's recognition of words may also be studied and compared with their reading ability in general. This has proven to be a valuable diagnostic test. 116 EXPERIMENTAL EDUCATION QUESTIONS AND TOPICS FOR DISCUSSION 1. Is there any resemblance between the perceptual process in reading and in the apprehension of the figures studied in Experiment No. 2? 2. Argue for or against the proposition that adult or developed recognition of words, or groups of words, implies previous analysis and synthesis of the elements of which it is composed. Because the adult does not notice all the details in his rapid recognition of a word, is his recognition like that of the child who is in the first stages of reading? ;8. Why are groups of digits not recognized as are groups of let- ters? What is present in the one and not in the other? 4. What individual dififerences are there with reference to the matter referred to in question 3 (the recognition of the ele- ments of a word)? 5. Is there evidence of distinct types ? If not what is the nature of the differences? /6. Are there disadvantages in very narrow or very broad scope of attention in word recognition ? What would be the effect of the two on reading ? n. Does the difference in the fusion of different kinds of words / indicate anything with regard to the possibility of increasing the scope of attention by training? REFERENCES Same as for Experiment No. 7, and in addition: Freeman, F. N. Psychology of the Common Branches, chap. 4. (Also on the other school subjects.) Judd, C. H. Genetic Psychology for Teachers, chap. vrti. Whipple, G. M. Manual of Mental and Physical Tests, vol. i, chap. vn. EXPERIMENTS WITH SCHOOL SUBJECTS 117 Experiment No. 11 efficiency of reading Problem. The practical application of the knowl- edge of the psychological nature of reading may be made by studying the conditions which determine the efficiency of reading. There are at least three elements to be taken into account in a complete determination of the efficiency of reading; namely, (1) the rate of reading, (2) the amount which is retained, and (3) the extent to which what is read suggests relevant trains of thought. The first two elements may be regarded as concerning efficiency of assimilation of what is read, and the third element as related to the ability to use what has been read. Previous experiments have dealt with the first two elements. On this account, and be- cause these are the elements most easily measured, they are made the subject of this experiment. Atten- tion is called to the fact, however, that final conclu- sions cannot be drawn until the third element also is studied. Material and method. In the present experiment, measurements will be made of the speed of reading and of the amount retained in reading two halves from a passage to be selected by the instructor. The passage should be ten pages or more in length and of uniform subject-matter. A division should be made in the mid- dle of the passage. The subject should first read the 118 EXPERIMENTAL EDUCATION first half of the passage at his ordinary rate, keeping the time with a stop-watch. He should then immedi- ately write what he can remember of it, trying to re- produce the ideas, but not necessarily the exact words. Let him then continue to read the passage, now reading as rapidly_as_ppssible consistent with a jjeai- grasp of thejneaning. When he has finished, he should measure the time required to read it and write what he remem- bers of it. Treatment of results. The individual reports should show clearly — first, how many words per second were read in each passage; and, secondly, what proportion of each passage was remembered. There are two methods of determining how much is remembered of a passage. We may designate these the verbal method and the idea method. Each individual in making up his report may use the verbal method. An exact form of the verbal method is to count the number of the particular words which were used in the original which are also to be foimd in the reproduction. The simpler procedure which may be used in this experiment con- sists in counting the number of words of the reproduc- tion which express ideas occurring in the original, and finding the ratio of their number to the number of words in the original. This is of course only a rough approximation, since the same idea may be expressed by one person in a few words and by another person in many. For the comparison of the two reproductions by the same person, however, it may serve. EXPERIMENTS WITH SCHOOL SUBJECTS 119 For the comparison of the amount which is repro- duced by the different individuals, which is the concern of the person who makes the general report, the idea method is fairer. By this method the original passage is divided into thought units. ^ The thought units may be made more or less comprehensive, and it is therefore desirable that one person grade all the papers. Hence this method is recommended for the general report only. The one essential is that the principle of division into units be the same throughout. A third method which is useful in some cases is to measure the correctness of answers to questions in the passage read. The problems which should be attacked by means of the data which are furnished by this experiment are in the main two: (1) What is the relation between the speed of reading and the amounts which may be repro- duced? and (2) Through what modifications of the mental process is the speed of reading increased? Each individual may gain light on these questions by ana- lyzing objective and introspective results from his own experiments. The general report should contain a summary of such facts as are brought out in the indi- vidual reports and should, in addition, study the corre- lation between speed and reproduction when different individuals are compared. To give an exact measure of such correlation re- 1 Cf. C. H. Judd, "Reading Tests"; Elementary School Teacher (l914),vol. 14, p. 371. 120 EXPERIMENTAL EDUCATION quires the application of a statistical method. A num- ber of such methods are described by Whipple in his Manual of Mental and Physical Tests, chap. iii. Method B, the Pearson Method adapted to Rank- Differences, may be used in this case or the more de- tailed tabular form of statement illustrated below may be employed. TABLE XI Words per No. of words No. of ideas Individr ual second reproduced reproduced Parti Pari II Pari I Pari II Pari I Pari II A 3.7 4.9 710 772 33 26 B 2.9 3.7 270 325 23 29 C 2.2 3.4 600 642 26 31 D 2.9 3.7 251 291 19 19 E 2.9 3.7 270 346 23 24 F 5.5 7.0 326 120 12 9 G 3.2 4.2 556 234 19 9 H 3.5 3.4 187 174 6 10 I 2.7 6.0 486 774 28 34 J 2.7 3.6 352 248 17 15 K 4.7 4.9 235 338 15 21 L 2.8 4.9 396 484 27 31 M 6.9 4.8 500 600 16 18 N 2.9 3.7 370 325 16 17 O 4.7 5.8 425 413 23 18 P 3.9 9.3 225 156 15 11 Q 5.3 5.2 359 513 39 29 B. 2.9 5.7 393 338 22 24 S 2.7 3.7 583 615 29 45 u 3.3 4.5 705 675 55 54 V 5.5 4.1 307 340 28 26 Total... 77.8 100.2 8406 8723 491 500 Average. 3.7 4.77 400.3 415.4 23.4 23.8 EXPERIMENTS WITH SCHOOL SUBJECTS 121 TABLE XII. THE DATA OF TABLE XI REARRANGED TO FACILITATE VARIOUS COMPARISONS Relation between rapidity of reading ike first and second passages Relation between rapidity of reading arid increase in rapidity A B C D a if 03 S 3 1 Is I ll .g I •3 1 1 ■S ■S 1 6.9 S.5 5.5 5.3 4.7 4.8 4.1 7.0 6,2 4.9 1 2 3 4 6 9.3 7.0 6.0 5.8 6.7 3.9 5.5 2.7 4.7 2.9 1 2 3 4 5 6,9 6.5 5.5 5.3 4.7 -2.1 -1.4 1.6 -0,1 0,2 1 2 3 4 5 6.4 3. 3 2.8 2.1 1.6 3 9 1 2.7 S 2 9 4 S>, R B 5.S Average 6.6 5.2 6.8 3.9 6.6 -0,4 3.0 3.8 6 4.7 8.9 3.7 3.5 3.3 S.2 6.S 9.3 4.9 3.4 4.5 4.2 6 7 8 9 10 11 6.2 4.9 4.9 4.9 4.8 4.5 5.3 4.7 3.7 2,8 6.9 3.3 6 7 8 9 10 11 4.7 3.9 3.7 3.5 3.3 3.2 1,1 6.4 1.2 -0.1 -1.2 1.0 6 7 8 9 10 11 1.2 1.2 1.2 1.1 1.0 1.0 3 7 7 S 3 8 2.2 9 4 7 10 3 « 11 9. 7 Average 3.7 6.3 4.9 4,5 3.7 2.0 1.1 3.3 Without 1 ... 3.7 4.6 4.9 4,4 3.7 0.9 12 2.9 2.9 2.9 2.9 2.9 3.7 3.7 3.7 3.7 6.7 12 13 14 IS 16 4.2 4.1 3.7 3,7 3.7 3.2 6.5 2.9 2.9 2,9 12 13 14 16 16 2.9 2.9 2.9 2.9 2.9 0.8 0.8 0.8 0,8 2,8 12 13 14 16 16 0,9 0,8 0,8 0,8 0,8 9. 7 IS ?, 9 14 9. 9 16 9, 9 16 9 9 Average 2.9 4.1 3.9 3.6 2.9 1,2 0,8 2.9 17 2.8 2.7 2.7 2.7 2.2 4.9 6.0 3.6 3.7 3.4 17 18 19 20 21 3.7 8.7 3.6 3.4 3.4 3.6 2.9 2.7 2.7 3.6 2.2 17 18 19 20 21 2.8 2.7 2.7 2.7 2.2 2.1 3.3 0.9 1,0 1,2 17 18 19 20 21 0.2 -0.1 -0.1 -1.4 -2.1 4 1 18 3 5 19 fi S 20 6.5 21 6 9 Average 2.6 4.3 3,8 2.6 1,7 -0.7 6.2 Final Average 3.70 4,77 4.6 4.77 3.70 3.70 1.06 1.2 3.7 Median 3.2 4.6 3,2 3.2 1.0 1.0 3.2 122 EXPERIMENTAL EDUCATION TABLE XII {continued) Relation between rapidity of reading and number of words reproduced in the first reading Relation between rajndity of reading and number of words reproduced in the second reading E F G H 1 1 1i li 1 1 2 3 4 5 g.S .1 1^ 1 2 3 4 5 ■si as 1^ 1.S Is 1 % CJ4 •S'S' II 1 Sag 1 ... 6.9 S.5 5.5 6.3 4.7 500 307 326 359 235 710 705 583 556 500 3.7 3.3 2.7 3.2 2.2 9.3 7.0 6.0 6.8 6.7 156 120 774 413 338 1 2 3 4 5 774 772 676 642 615 6 2 4 9 3 4.5 4 3.4 5 3.7 Average 6.6 345 611 3.0 6.8 360 696 4.6 6 4.7 3.9 3.7 3.5 S.3 3.2 42S 225 710 187 70S 556 6 7 8 9 10 11 500 486 425 396 393 370 6.9 2.7 4.7 2.8 2.9 2,9 6 7 8 9 10 11 6.2 4.9 4.9 4.9 4.8 4.6 513 772 338 484 600 675 6 7 8 9 10 11 600 513 484 413 346 340 4.8 7 ... 5.2 8 4 9 9 6 8 10 S 7 11 ... 4.1 Average 3.7 468 428 3.8 4.9 664 449 4.8 12 2.9 2.9 2.9 2.9 2.9 270 251 270 370 393 12 13 14 15 16 859 352 326 307 270 6.3 2.7 6.5 6.6 2.9 12 13 14 15 16 4.2 4.1 3.7 3.7 3.7 234 840 325 291 346 12 13 14 15 16 338 388 326 325 291 4 9 13 5 7 14 R 7 15 3 7 16 S 7 Average 2.9 311 323 4.3 3.9 307 322 4.3 17 2.8 2.7 2.7 2.7 2.2 398 486 352 583 500 17 18 19 20 21 270 251 235 225 187 2.9 2.9 4.7 3.9 3.6 17 18 19 20 21 3.7 3.7 3.6 3.4 3.4 325 615 248 642 174 17 18 19 20 21 248 234 174 156 120 3.6 18 4 2 19 3 4 20 21 9.3 7 Average 2.6 464 234 3,6 3.6 4.8 401 186 6.5 Final average S.7 400 400 3.7 415 416 4.8 Median 3.2 370 370 3,2 4.S 340 340 4.6 EXPERIMENTS WITH SCHOOL SUBJECTS 123 TABLE XII {continued) RelaHon between increase in rajndity and increase or decrease in words and ideas reproduced Relation between rate of reading ^ gain or loss in rate and gain or loss in reproduction I J 1 g .e ^ S .1 } .3 »5 •S 1 1 S.4 3.3 2.8 2.1 1.5 - 69 288 - 5S - 88 -206 - i 6 2 4 - S 3.9 2.7 2.9 2.8 5.5 21. 6. 20. 19. 17. 6.9 5.6 6.5 6.3 4.7 -2.1 1.6 -1.4 -0.1 .2 100 -206 38 154 - 12 2 g S 4 10 5 5 Average 3.0 9 1 3.6 6.6 - .4 14 - 4 6 1.2 1.2 1.2 1.1 1.0 1.0 62 - 30 142 - 12 -S22 32 - 7 - 1 B - 5 -10 16 3.7 8.3 2.2 4.7 8.2 2.7 9. 1. 6. 18. 7. 10. 4.7 3.9 8.7 3,5 8.3 3.2 1.1 6.4 1.2 -0.1 -1.2 1.0 103 - 69 - 62 - 13 - 80 -322 7 8 _ 7 9 10 10 — 1 11 —10 Average 1.1 - 21 - 0.33 3.3 8.7 1.6 - 47 - 1 12 .9 .8 .8 .8 .8 -104 55 40 76 - 45 - 2 6 i" 1 2.7 2.9 2.9 2.9 2.9 3. 13. 14. 15. 16. 2.9 2.9 2.9 2.9 2.9 2.8 .8 .8 .8 .8 - 66 66 40 76 - 45 2 13 6 14 15 1 16 1 Average .8 4 1 2.9 2.9 1.2 14 2.6 17 .... .2 - .1 - .1 -1.4 -2.1 103 - 13 154 33 100 6 10 -10 - 2 2 4.7 3.5 6.3 5.5 6.9 4. 2. 11. 12. 8. 2.8 2.7 2.7 2.7 2.2 2.1 3.3 1.0 .9 1.2 88 288 82 -104 142 4 18 6 19 ....... . 16 20 - 2 21 6 Average - .7 75 1 6.2 2.6 1.7 89 6 Final average 1.2 15 - .71 3.7 3.7 1.1 16 .75 Median 1.0 83 1 3.2 8.2 1.0 33 1 124 EXPERIMENTAL EDUCATION Results of the experiment. The summary results of this experiment are presented in Table XI. These results have been reformulated to facilitate comparison in Table XII. The data are presented in the general table in order that one may follow the record of any individual throughout the table. In Table XII the individuals are rearranged in each section so as to put the scores into regularly descending order. The first comparison to be made is between the ra- pidity of reading in the first and second halves of the selections. The increase in rapidity as a result of the effort to read rapidly is large for the class as a whole, being 27 per cent if we use the averages, and 41 per cent if we use the medians. The more particular question which is presented by our results concerns the relationship between the standing of the individual subjects in rapidity of read- ing of the two halves of the selections. Are those who read rapidly in the reading at the ordinary rate the most rapid readers when they try to increase the speed? This is the type of problem which involves the calcula- tion of correlation. If the order of the scores in the two series is the same, or if there is any more correspon- dence than would be present by pure chance, the cor- relation is positive. If the order is reversed, partly or wholly, the correlation is negative. The more precise methods of calculating or representing the degree of correlation will be illustrated in Experiment No. 16. We shall be content here with the arrangement of the EXPERIMENTS WITH SCHOOL SUBJECTS 125 data in such a way as to facilitate an inspection of them, and the detection of the outstanding facts of correlation. Sections A and B present the data regarding the cor- relation between the speed of reading of the class in the two parts of the selection. In Section A the individu- als are arranged in the order of the rapidity with which they read the first half, and in Section B in the order of rapidity of reading the second half. As a very rough method of estimating the correlation the series are divided into four approximately equal groups, and the average is given for each group. The averages, of course, decrease as one goes down the column in the left-hand column in each section. If the averages of the right-hand column also decrease in the same order there is probably some correlation. Inspection of the individual scores makes possible more detailed state- ments. It appears that there is some correlation between the rapidity of the first and second readings. In gen- eral, those who read rapidly in the first part read rap- idly also in the second part. Those who read most rapidly in the second part, however, were not as rapid in the first part as were the second group. Those who read most rapidly the first time apparently were read- ing nearly at their maximum and did not increase their speed (with one exception), but actually fell oflf in the second reading. All the other groups, as appears from Section C, increased their speed materially. But while 126 EXPERIMENTAL EDUCATION some of the slow readers in the first half of the test read rapidly in the second half, those who were slow when they read under pressure were in the main slow when they read at their natural rate. We may conclude from these facts that those who read most rapidly in the test for natural reading were reading about as fast as they could. Some of those who read more slowly were actu- ally capable of more rapid reading than the most rapid readers in the first test; and some of those who were toward the bottom of the class in the first test ranked in the top quarter in the second. Some of the slow readers in the first test, however, while capable of some increase in speed, remained at the bottom; and very rarely did a rapid reader in the first test fall toward the bottom in the second. Sections C and D indicate that the slower readers in the first test, while they in general remained below the rapid readers when they intentionally speeded up, yet gained more, absolutely, and very much more rela- tively, than their rapid companions. These facts taken together indicate that, so far as a single test can be relied upon, the slow readers were slow partly because of some native or deep-seated acquired difference, but that the difference was in most cases greater than necessary since it could be overcome by a little effort. From a study of Sections E to H, which show the relationship between rapidity of reading and the amount which is reproduced it appears that there is a diversity of cases. If we examine Sections E and G we EXPERIMENTS WITH SCHOOL SUBJECTS 127 see that some rapid readers remember a large amoimt of what they read — for example. No. 1 in Section E and No. 3 in Section G. It also appears from the same sections that some of the slow readers — as Nos. 18, 20, and 21 in Section E, and Nos. 18 and 20 in Section G — reproduce a large amount. In the same manner it may be shown that there are both fast readers and slow readers who remembered little of what they read. Columns F and H show the converse of these facts. While there are individual cases which show a cor- respondence between speed and reproduction, and other cases in which there is a wide discrepancy be- tween the two elements, there may still be in general a positive or a negative correlation. If there is a large amount of correlation, either positive or negative, it should be indicated by the averages of the groups in the second column of each section. There is clearly no strongly marked positive correlation, but, on the other hand, some negative correlation. In each case the five most rapid readers reproduce less than the five slowest, and the five who reproduce most read more slowly than the five who reproduce least. These facts may be summarized in the following statements : — Some individuals read rapidly without sacrifice to comprehension, while others attain speed at the ex- pense of comprehension. The latter are slightly more numerous than the former. Some individuals read slowly without a correspond- 128 EXPERIMENTAL EDUCATION ing gain in comprehension, while others sacrifice speed to an unusually high degree of comprehension. The latter are slightly more numerous than the former. Section I presents the data for the comparison of the rate of increase or decrease in speed from the first to the second reading, with the increase or decrease in the number of words or ideas reproduced. There is a wide diversity among individuals and no clear general re- sult. The most that can be said in general is that the increase in speed for the class as a whole, which amounts to twelve words per second on the average, or about 32 per cent, does not result in any diminution of the amount reproduced. This appears to be a clear gain for the class as a whole. Among individuals there is a great diversity. Indi- vidual No. 2, for example, increased 3.3 words per second in rate of reading, and gained 288 words and 6 ideas in reproduction. This person read slowly in the first reading, the rate being 2.7 words per second. An- other individual. No. 17, who read at approximately the same rate during the two readings (4.7 and 4.9 respectively) increased the amount reproduced by 103 words and 6 ideas. On the other hand. No. 5, who gained 1.5 words in rate of reading, lost by 206 words and 3 ideas in amount reproduced. This person was a fast reader, reading at the rate of 5.5 words, and probably made too great an effort at speed in the second reading. These cases suggest that by taking into considera- EXPERIMENTS WITH SCHOOL SUBJECTS 129 tion the rate of reading as well as the gain, a relation- ship with gain or loss in reproduction may be discov- ered. This proves to be the case. In order to display the relationship Section J of the table is constructed, in which the individuals are classified on the basis of rate of reading in the first test. Of the five most rapid readers the only one who in- creased materially in speed of reading, No. 5, lost heavily in the amount reproduced. The other four, who either lost in speed or gained but slightly, either gained in words or ideas or lost but very slightly. Of the five slow readers, on the other hand, four gained materially in speed as well as in the amoimt repro- duced in both words and ideas. The slowest readers were benefited in every way by their increase in speed. In the middle half of the whole group the result is not so uniform, but in general the slower readers of the third group gained in reproduction as well as in speed, while the faster readers of the second group gained in speed but lost in reproduction. The results of this comparison between the increase in speed and the increase in reproduction, taken to- gether with the relationship between the absolute rate of reading and the amount reproduced, warrant a statement which, while there are exceptions, holds in general so far as this class is concerned. This state- ment is as follows: The most rapid readers in the majority of cases sacrifice comprehension to speed, and the slowest readers sacrifice speed to comprehension. 130 EXPERIMENTAL EDUCATION Furthermore the slow readers can increase the speed materially without loss in comprehension, or even with a gain, while an increase in speed among the rapid readers usually results in a loss in comprehension. Extension of the experiment. These conclusions demand extension and verification through the collec- tion of results from larger numbers. One very useful extension could be made by giving several tests with different kinds of subject-matter to the same group of individuals. A single test is never an adequate basis on which to measure the ability of an individual. It would be well to make a number of tests with a group of per- sons without saying anything about speed at first, in order to obtain a reliable measure of reading ability at the ordinary rate. The next step would be to give a whole series of texts to be read rapidly in order to find out the amount of improvement in speed of which the various individuals were capable, the amount of prac- tice to bring each one up to his approximate maximum, and the final effect of the increase in rate on the com- prehension of the various individuals. A less extensive experiment than the foregoing might well be made by repeating the experiment with various kinds of subject-matter, as essays, fiction, and poetry. Another variation which usually gives very good results consists in having a limited number of individuals — even so few as one will do — carry on a practice experiment for the ptirpose of increasing the rate of reading. EXPERIMENTS WITH SCHOOL SUBJECTS 131 QUESTIONS AND TOPICS FOR DISCUSSION In addition to the specific problems mentioned above, the follow- ing questions may be discussed : — 1. What seems in your experience to be the effect of the following factors in determining the rate of reading: eye-movement hab- its; the amount of inner articulation used; the rapidity of appre- hension of meaning; attention to word characteristics (diction, shades of meaning, sound, characteristics as used in alliteration, onomatopoeia, etc.), melody, balance of sentence structure? 2. Should you say that the best speed of reading of different sorts of subject-matter, e.g. scientific prose, essays, novels, poetry, is the same? What considerations determine the best speed for these different kinds of material? 3. How may speed of reading be increased? 4. Has rapid reading any conceivable advantage beside economy of time ? 6. Should speed of reading be uniform ? 6. How may one strike the proper balance between speed and apprehension of meaning ? 7. Propose means of improving the apprehension of meaning. REFERENCES Freeman, F. N. Psychology of the Common Branches, chap. 4. Judd, C. H. Measuring the Work of the Public Schools. Cleveland Foundation. Starch, D. "The Measurement of EfiBciency in Reading"; in Journal of Educational Psychology (1915), vol. 6, pp. 1-24. Thomdike, E. L. "The Measurement of Ability in Reading''; in Teachers College Record (1914), vol. 15, pp. 1-67. Waldo, K. D. "Tests in Reading in the Sycamore Schools"; in Elementary School Journal (1915), vol. 16, pp. 861-68. 132 EXPEKIMENTAL EDUCATION Experiment No. 12 apprehension of number Problem. The abstract idea of number is the out- growth of various forms of concrete experience, and in .most if not all cases it never becomes entirely independ- ent of some sort of imagery which reflects concrete experience in visual or mimetic terms. The study of visual experiences which form the basis of number apprehension, and which exhibit it in simple and more elaborate forms, will therefore give some insight into the means by which the idea of number is developed. There are a variety of forms of concrete experience through which the idea of number may be developed. One is counting, which consists in giving attention to a series of objects in succession, at the beginning guid- ing the attention by pointing, handling, objects, etc. and designating each object in the series by a number name. Another is the division of an object — a length, surface, or solid — into equal parts, and designating the sum of the parts by a number name. The third form of experience, which will constitute the basis of this experiment, consists in the simultaneous apprehen- sion of a group of objects, and the designation of the group by a number name. What we may call the perception of number exists in each of these cases prior to the use of number names, and forms the basis of the idea which is represented by the verbal expression. EXPERIMENTS WITH SCHOOL SUBJECTS 133 Our problem in this experiment is to study the effect of grouping upon number apprehension. Material and method. The most convenient method of studying the simultaneous apprehension of a group of objects is to present the group to view for a short space of time by the brief exposure or tachistocopic method. We shall use groups of dots placed upon white cards and presented by means of the faU exposure apparatus. In order to test the effect of grouping, one series of cards contains dots arranged in a straight hori- zontal row, with equal spaces between them; another series contains dots arranged in irregular grouping; and the other series contains dots in certain regular forms of grouping. The stimuli used are reproduced in the Appendix, which see. The cards of each series should be presented in irreg- ular order, so that the subject may not be able to anti- cipate what the number about to be presented may be. As each card is reached in the series it should be pre- sented once (after warning signal), and the subject should make a record of his judgment of the number of spots, of their arrangement, and any introspections he may make. Each judgment may afterward be identi- fied and its correctness determined by labeling it with the number of the series and its number in the sequence of presentations in the series. After all the series have been presented once, they may be presented a second time, and each subject's score may be based upon the second trial. Experimenter and subject may exchange 134 EXPERIMENTAL EDUCATION places each time after the completion of the presenta- tion of the whole series of cards. Treatment of results. The numerical results of this experiment may be simply expressed. Each individual should report his judgments in the two trials upon each number of each series. For this purpose the num- bers should be arranged in consecutive order. He should then determine and record the highest number of each series which was correctly judged in the second trial. Furthermore, he should classify the errors in each series and in the whole number of judgments into underestimations and overestimations, and calculate the proportion of each. The general report should contain a table and chart based upon the percentage of correct judgments (based upon the second judgment in each case) which were passed by the class as a whole upon each stimidus card. The chart and table should exhibit a comparison of the percentage correctness of the judgments of the same numbers in the diflferent series. TABLE XIII. PERCENTAGE OF CORRECT JUDGMENTS FROM A GROUP OF TWENTY INDIVIDUALS UPON EACH NUMBER OF THE VARIOUS SERIES Number 3 4 5 6 7 8 9 10 11 12 13 14 15 Series I.... 80 45 40 40 35 25 II.... 100 85 75 70 60 45 45 III.... 85 80 80 75 60 65 55 55 IV.... 90 85 85 70 45 60b SO 25 40 v.... 95 75 70 100 70 60 90 55 30 45 EXPERIMENTS WITH SCHOOL SUBJECTS 135 Results of the Ejcperiment. Table XIII and Chart V give the data for the comparison of the apprehension of the various sorts of arrangement of the dots. The facts may be most readily inspected in the chart. It is Percentage 80 Number exposed 3 4 6 6 7 8 9 10 11 12 13 14 16 \ " ■" " ~ - V \ / \ , s s . » t \ ,'■ s \ ! \ \ ^ ■ N, \ \ s , y \ \ ^ \ \ \ s, \ \ 'i s \ X ,1 \ \ \ \ A \ \ [ \^ 1 ■Ix. \ \ / I \ V -I A A T N ,1- » K t / \ \ 1 ', s \ ' / ^l '- •J -J Cbaet V. GRAPHIC REPRESENTATION OF THE PERCENTAGE QV CORRECT JUDGMENTS GIVEN IN TABLE XIH I. Irregular arrangement II, Horizontal row III. Grouping by fours IV. Grouping by threes V. Grouping by fives apparent at a glance that there is a radical difference in the mode of apprehension of the dots in irregular grouping, and in grouping by fives, for instance. Se- ries I and Series V. By comparison of the individual series and by a study of the introspections the effect of the various forms of grouping may be made out. 136 EXPEKIMENTAL EDUCATION The significance of the results should be studied with reference to the light they throw upon the span of attention and the mode of organization which is pro- duced by grouping. The apprehension of objects in groups is particularly significant as an indication of the nature of much of our recognition of number. Question 4 suggests one application of the facts. What are others? TABLE XIV. PERCENTAGE OF ERRORS CONSISTING IN OVERESTIMATIONS IN THE VARIOUS NUM- BERS OF THE DIFFERENT SERIES (The numbers in italics represent the total number of errors) Number 3 4 5 6 7 8 9 10 11 12 13 14 15 Series I... 100 11 64 42 IS 25 13 46 15 13 II... 3 5 6 8 11 11 100 40 50 50 9 9 III... 3 4 ^ 5 8 r 9 9 100 75 75 20 50 45 9 22 IV... s 3 3 6 11 8 la 100 100 66 83 73 38 14 i4 8 V... 1 6 6 6 8 2 9 i^ 11 100 100 17 33 75 50 9 14 18 Table XIV presents the facts regarding the relation of overestimation and underestimation in the appre- hension of the various sized numbers of the diflPerent series. The table shows that in all the series the errors in the smaller numbers consist chiefly in overestima- tions, while in the case of the larger numbers the errors EXPERIMENTS WITH SCHOOL SUBJECTS 137 are underestimations. What is the cause of this marked difference? Is it an evidence that an error in the apprehension of a smaller number of objects is of a different sort from an error in the apprehension of a larger number, or is it due to the same cause which in some way produces a different result under the two conditions? Introspections should be made use of in discussing this question. All the hypotheses possible should be presented, and their probability discussed. TABLE XV. INDIVIDUAL DIFFERENCES IN PER- CENTAGE OF CORRECT JUDGMENTS Individual Percentage of cor- rect judgments Individual Percentage of cor- rect judgments A 10.0 K 67.5 B 45.0+ L 70.0+ C 45.0 M 70.0 D 47.5 N 72.6 E 55.0 O 72.5 F 57.5 P 75.0 G 62.5+ Q 75.0= H 62.5 R 77.8+ I 65.0 = S 80.0 J 67.5 T 82.5 Table XV shows individual scores in the percentage of correct judgments made. The significance of such marked individual differences as are represented in this and other tables in these is one of the important prob- lems of educational psychology. In such an experi- ment as this it is likely to occur to the student that purely sensory differences are likely to be responsible 138 EXPEEIMENTAL EDUCATION for the wide divergence in the results, but there seems no good ground for this conclusion. This hypothesis may be tested by comparing the records of those who wear glasses with those of those who do not. There is some indication of a difference in the mode of appre- hension in the difference in the percentages of over- and underestimations. In the majority of cases the total number of underestimations exceeds the over- estimations, but in the case of four individuals, indi- cated in the table by the plus sign, the reverse is the case; and in the case of two, indicated by the equality sign, the over- and underestimations are equal in number. These divergent cases, however, are regularly distributed among those of high and low percentages of correct judgments. It is very desirable to be able to determine the sig- nificance of such facts of individual difference as this, both for schoolroom practice and for the diagnosis of abilities in vocatioijal guidance. For a discussion of some of the attempts which have been made in the latter sphere see Miinsterberg.^ Individual differences in the fundamental forms of number ability have been studied particularly by S. A. Courtis. See his article in the report of the New York School Inquiry.^ Extension of the experiment. A valuable extension ' M.\insteiheig, H., Psychology and Industrial Efficiency. Boston: Houghton Mifflin Co. 1913. ' Courtis, S. A., "Report on the Courtis Tests in Arithmetic"; in Report of Committee on School Inquiry, City of New York (1911- 13), vol. 1, pp. 389-546. EXPERIMENTS WITH SCHOOL SUBJECTS 139 of this experiment consists in making the experiment with children of various ages, and noting the type of recognition in which the older children and adults show most development. An attempt may also be made to determine whether individual differences among children reflect their ability in number work, or in any other phase of their school work. QUESTIONS AND TOPICS FOR DISCUSSION The student may judge which questions can be answered only in the general report. 1. What is the average scope of attention as measured by the number of different unorganized units which can be appre- hended simultaneously? Within what limits does it vary? 2. How does grouping (subjective or objective) enable one to judge correctly a number of objects beyond the scope of attention? 3. Which do you think is more affected by education, the scope of attention or the ability to apprehend groups? (If opportunity offers, thb question could well be tested.) 4. Does the experience obtained in this experiment suggest any means by which we may form the idea of numbers larger than can readily be grasped concretely? Does the decimal system suggest any analogy? 5. What possibilities exist in the procedure used in this experi- ment for the teaching of number? 6. Point out other kinds of experience in which the organization of the material increases the amount which can be appre- hended as a unit. REFERENCES Freeman, F. N. "Grouped Objects as a Concrete Basis for the Number Idea"; in Elementary School Teacher (1912), vol. 12, pp. 306-14. Judd, C. H. Genetic Psychology for Teachers, chap. ix. McCIellan, J. A., and Dewey, J. Psychology of Number. CHAPTER IV TESTS There are two senses which are important for the child's education and mental development, and in respect to which there are a considerable number of children in whom there is marked deviation from the normal; namely, vision and hearing. The purpose of these tests is the detection of serious deviations from the normal, and the rough determination of the charac- ter of the defect. The aim is not to determine with precision the amount of the defect nor to equip the student to prescribe remedies other than those which are included in the duty of the school administrator. TESTS 141 EXPEBIMENT No. 13 TESTS OF VISUAL DEFECTS Problem. The defects which are to be tested in this experiment are myopia (near-sightedness), hyperopia (far-sightedness), astigmatism (lack of clear focusing of the image on the retina), and heterophoria (lack of muscular balance, which produces a tendency of the eyes to cross, squint, etc). a. The first test is for the purpose of detecting the presence of myopia or hyperopia. Myopia. This is usually caused by too great length of eyeball from front to back, as a result of which the image of an object at a distance is brought to a focus in front of the retina. Hyperopia is caused by the op- posite fault, as a result of which the image would come to a focus back of the retina if the crystalline lens of the eye were at rest. Myopia cannot be corrected by the action of the crystalline lens, because this would necessitate that the lens become flatter than it is when at rest. The lens flattens, however, only through the "Tjontraetien of-^ muscle which is constant in its ten- sion, and which cannot be relaxed or tightened to suit the varying needs of accommodation; and hence there is no means of further decreasing its curvature. For this reason the defect is easily discovered by the inabil- ity of the subject to clearly distinguish objects as far away as can be seen by the normal eye, and by his 142 EXPERIMENTAL EDUCATION ability to see objects at a greater distance by the use of a lens which brings the light to a focus farther back; that is, a concave lens. H3rperopia. This, on the other hand, is detected (provided the ciliary muscle which controls the crys- talUne lens has not been paralyzed by a drug) in a somewhat more roundabout way. After the analogy of myopia, we might assume that the, far-sighted indi- vidual could not distinguish objects close by, but this assumption would be incorrect, for the thickening of the lens which is necessary to bring it to a focus upon objects near at hand is produced by the contraction of the muscle of accommodation — the ciliary muscle — and in the hyperopic eye it is only necessary that this contraction be more vigorous than in the normal eye. A further consequence of the defect is that even when the lens is accommodated for distant objects the ciliary muscle is not at rest, as in the normal eye, but must be more or less strongly contracted. It is this fact which is employed to detect hyperopia. If the normal eye is accommodated for a distant object and a convex lens is placed before the eye, producing in effect an accom- modation for a point nearer by, the lens cannot adjust itself by further flattening, since it is already at rest; and the object as a consequence appears blurred. In the hyperopic eye, on the other hand, the lens is still somewhat thickened by the ciliary muscle, and this by further relaxing can compensate for the effect of the artificial lens, with the result that the object remains TESTS 143 clear. In a word, then, if the addition of a convex lens does not blur the vision of an object set at the maxi- mum distance, the eye is hyperopic. Material and method. The test consists merely in applying these general principles. In doing so, it is, of course, necessary to" examine one eye at a time, and to keep the other eye covered without pressing upon it. Placing a blank disk in the trial frame, is a convenient method of doing this. The McCallie vision test cards ^ are recommended, because their order can be changed at will and ordy one letteS" is shown at a time, thus making an accurate record possible of what is and what is not correctly perceivedv / Detailed directions. Eequire the subject to sit twenty feet from the spot where the cards lie face down upon a table. Cover one eye. Present the first card with the small letter at the top and ask the subject to name it. Record his answer, and continue with the other cards until ten have been presented, or until it becomes evident that the subject cannot read four out gfjfiye of the letters at that distance. If he has read correctly eight out of ten cards, there is no considerable myopia, and the test should be made for hyperopia by placing/the convex lens (+.75) in the trial frame. If the subject can still read the smallest letters, there is evidence of hyperopia. If they appear blurred, the ^ These cards may be purchased from Edwin Fitzgeorge, agent. Box 67, Trenton, N. J. A set for literates contains letters. Another set may be used with ilUterates. 144 EXPERIMENTAL EDUCATION eye is approximately emmetropic — that is, neither hyperopia nor myopic. If the subject cannot read eight out of ten of the smallest letters, the next larger letters should be pre- scribed; and if he fails to read these, the next larger; and so on. If he fails to read the largest letters, the distance should be reduced by two feet at a time until he can do so. The trial should be made now with the concave lens (—.75) to determine whether vision is improved thereby. If so, the eye is myopic. The results may be conveniently recorded by means of the following formula: The acuity of vision is ex- pressed by the ratio of the greatest distance from which a particular size of letter can be read to the greatest distance at which it can be read by a normal eye. Thus, if the letters marked forty are the smallest which the subject can read at a distance of twenty feet, his acuity is recorded as 20/40. It should further be recorded whether the eye is emmetropic, hyperopic, or myopic. The same procedure should be gone through with the other eye. b. The second test is for the purpose of detecting the presence of astigmatism. Astigmatism. This defect is caused by an unequal curvature of one of the refracting media of the eye in its various meridians. The seat of the trouble is usually the cornea. As a result the rays of light from any given point will not reach a focus at a point, but those which meet the cornea in one meridian will be focused at one TESTS 145 point and those which meet it in another meridian will be focused at another point. The retina may be situ- ated at the focus of one or the other of these meridians, or at a point different from either, and there are in con- sequence a variety of possibilities with regard to the retinal image. The common result, however, is that there is an elongation of the image in some particular direction due to the fact that the focal distance of the cornea in this direction is greater or less than that of the rest of the cornea. Any line, then, which is in any other than this direction of unequal curvature will appear blurred because of the elongation of the image in a lateral direction. In the case of a line which is parallel to the direction of unequal curvature, on the other hand, the elongations all fall within the line itself and hence do not produce blurring. The common method of detecting astigmatism, therefore, is to pre- sent a card which contains lines in various directions radiating from a center. If the lines in one direction appear distinct, while the rest are blurred, there is evi- dence of astigmatism. If astigmatism is detected by this means, the defect is sufficient to require correction. Oculists usually make a more exact diagnosis by means of trial cylindrical lenses and the opthalmoscope. The test should be made as follows : Place the chart at a distance of twenty feet from the subject and test one eye at a time. If myopia or hyperopia exist, insert in the trial frame the lens which partially corrects the defect. Ask the subject whether the line in one direc- 146 EXPERIMENTAL EDUCATION tion appears blacker and more distinct than the ot^iers. If the answer is aflBrmative, the correctness of the judg- ment may be confirmed by turning the head so that the one eye is raised and the other lowered. The position of the distinct lines should shift correspondingly. If an aflSrmative answer is thus confirmed, the defect may be recorded as astigmatism in the axis of the lines which appear distinct. c. The third test. This is for the purpose of detect- ing heterophoria, or the lack of muscular balance. Heterophoria. This is caused by an unequal tension in the pairs of muscles which turn the eye out and in or up and down, or a combination of these two conditions. This inequality of tension may not produce an actual displacement of the line of vision of the eye, since the motive of clear vision may lead to an extra innervation of the muscle opposing the muscle which possesses un- due tension. This, however, produces a condition of strain which may have serious effects upon the nervous system. The test for heterophoria is based upon the fact that if the motive for directing the eyes toward the same point is not present, the defect becomes manifest. This may be done by placing a lens or other medium before one eye so that the image in that eye is distorted and is not recognized as being produced by the object which is perceived through the other eye. The Stevens stenopaic lens* is recommended for this purpose, since 1 The stenopaic lens should be tested to see whether it is accurate. If it is not, as is likely to happen, the Maddox rods may be used and TESTS 147 it reveals the direction and amount of displacement by a single determination. This is a convex lens of short focus covered except for a small opening in the center. When this lens is placed with the opening close to and in front of the pupil, a source of light, as a candle, is seen as a circular spot of diffused light. If the muscular balance of the eyes is normal, the image of the candle as seen through the other eye should appear in the center of this spot. If it does not, there is heterophoria. Proceed as follows. Place the stenopaic lens before one eye. Place a lighted candle on a level with the eyes and twenty feet away. Direct the subject to close his eyes and immediately upon opening them and looking at the candle to say whether it appears in the middle of the circle of light; or if not, what its position is. Repeat the test with the other eye. The results of the test should be recorded for each eye separately. The record should state whether there is heterophoria; and if there is, in what direction the lack of muscular balance exists. Results of the experiment. The record of the test of a group of twenty-two individuals for visual acuity is shown in Table XVI. In most cases the record is con- sistent. For convenience in inspecting the table the cases in which a ratio less than 20/20 indicates a de- fect in acuity are underlined. In two cases there is a defect indicated in both eyes, and in five cases in one the test made with the instrument in both the vertical and the hori- zontal positions. 148 EXPERIMENTAL EDUCATION TABLE XVI. RECORD OF THE TESTS FOR VISUAL ACUITY OF TWENTY-TWO INDIVIDUALS Ratio OS Letlera recog- nized Large letters re- cognized Letlera recognized with lensee Convex % 1-5 Concave A. B. C. D. E. F. G. H. I.. J.. K. L. M. N. O. P. Q. R. S. T. U. V. 20/20 20/20 20/30 20/20 20/30 20/20 20/20 20/30 20/20 20/20 20/20 20/20 20/30 20/20 20/20 20/30 20/20 20/20 20/20 20/20 20/20 20/20 20/20 20/20 20/20 20/30 20/20 20/20 20/20 20/30 20/20 20/20 20/20 20/30 20/20 20/20 20/20 20/20 20/20 20/20 20/60 20/20 20/20 9* 8* 8t 8Bt 4B 8B.L. 5B OB 10 10 6 10 not so good 3B 10 B.E. 7B 8B 10 neither lens helps. OB 9 9B 6B 10 8 9 B 8 6B 7 10 no data 10 6 5 * Second size. t Fourth size. J Blurred, TESTS 149 eye only. Usually when eight or more of the smallest letters can be recognized with the naked eye the recog- nition is impaired by the use of either lens. In the case of Subjects D (R.E.), G, and U, however, the use of the convex lens either produces improvement or fails to impair vision. What does this indicate? In the case of four individuals vision was improved in one or both eyes by the concave lens. What does this indicate? One case, that of M, appears, so far as this test can be relied upon, to be improved by neither lens. This may point to astigmatism, or some other form of de- fect, such as cataract or retinal defect. From reference to the next table no astigmatism is apparent, though expert examination would be necessary to render judgment on this point certain. The record of the test in astigmatism, Table XVII, indicates a rather large proportion, 50 per cent of de- fective cases in this rather rough method of determina- tion. The results of this method must be taken with considerable allowance for error in observation. The results of heterophoria are presented not merely to illustrate the proportion of cases in which hetero- phoria is to be expected, but also to indicate that care is needed in interpreting the objective results. Take Subject D, for instance. He finds heterophoria in one eye and not in the other. Is this possible by the method used? The test depends on the comparative positions of the two eyes. Or take Subjects E, G, I, K, M, O, U, and V. They find, according to their report, esophoria in 150 EXPEEIMENTAL EDUCATION TABLE XVII. RECORD OF THE TESTS FOR ASTIG- MATISM AND HETEROPHORIA OF TWENTY-TWO INDIVIDUALS + indicates presence, and — absence of defect Astigmatism Heterophoria in horizontal plane Subject R.E. L.E. R.E. L.E. H' Es-' Ex-' HI Es-' Ex^ A B C D E F G H I J K L M N P Q R S T U V + + + + + + + no + + + + + + + + + + data + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + 1 Heterophoria. ^ Esophoria, or displacement inward. > Exophoria, or displacement outward. one eye and exophoria in the other. Is it possible to get this result by this method? In order to answer these questions, and the further TESTS 151 question, — what does the displacement of the line to the right or left mean with the lens before each eye with reference to the relative position of the eyes? — the student should draw a diagram of the eyes, the source of light and the various possible positions of the image of the light and of the line. Extensions of the experiment. Possible extensions of this experiment will be described at the end of Ex- periment XIV. REFERENCES For an account of the frequency of these defects, see G. M. Whip- ple, Manual of Mental and Physical Tests, chap. vi. For a full de- scription of the various defects and their physiological basis, see H. Eulenberg and T. Bach, Schulgesundeheitslehre (1900), vol. 1, pp. 748#. For questions and discussions, see the end of the next experiment. 152 EXPERIMENTAL EDUCATION Experiment No. 14 tests op auditory acuity The most serious defect of hearing is the inability to hear sounds of an intensity (loudness) suflBcient to be heard by the individual with normal hearing. The manner of testing auditory acuity is simple in principle, but the conduct of the test is attended with difiBculties. It is not a simple matter to produce a series of sounds in regularly ascending or descending grades of intensity and of standard intensity. One means at hand is to vary the distance of the source of sound from the ear; but this method is attended with the complicating factor of reflecting walls, and ordinarily by the disturb- ing presence of other sounds in the neighborhood. This is the method of the watch test or the whisper test. To overcome these difficulties, the sound may be applied to the ear and its intensity varied in regular steps. This is the method of the audiometer. A convenient general method of procedure, whatever the form of stimulus which is used, is a combination of the method of right and wrong cases with the method of varying the stimulus. This method consists in de- termining the intensity of the stimulus at which the subject will give a correct answer in eight out of ten judgments. The detailed procedure with the watch test, the whisper test, and the audiometer is as follows: — TESTS 153 Directions a. Watch test. Blindfold the subject, close one ear with the finger or a plug. Hold the watch opposite the open ear well within hearing distance and slowly move it away, requiring the subject to say at intervals whether or not he hears it. When the point has been reached at which the subject reports that he no longer hears the watch, mark the spot on the floor with chalk. Begin beyond the point at which the watch can be heard and move it slowly toward the subject until he clearly hears it and mark as before. Then tell the sub- ject that when the watch is presented it will always be at the same place but that, in approximately half the cases in which he is asked to judge, there will be no stimulus whatever. Then hold the watch midway between the two marks and proceed by distributing the cases in which the watch is and is not presented irregularly, but in about equal number. If the answers are correct in eight out of ten cases, measure the dis- tance of the place of stimulation from the ear and re- cord it as the threshold. If the answers are correct in smaller ratio than eight to ten, move the watch nearer and elicit another set of judgments, and so on until a place is found where the ratio is as required. Proceed in the same way to test the other ear. b. Whisper test. The disadvantage of the watch test (besides variation in the loudness of tick of diflfer- ent watches, and the consequent absence of an abso- 154 EXPERIMENTAL EDUCATION lute standard) is that the sound is not one which one is accustomed to Hsten for, and is one which it is very easy to imagine one hears when one does not. Further- more, there is no means of knowing certainly whether the subject actually does hear the sound; and finally, the abiUty to hear the tick of a watch does not always correspond entirely with the abUity to hear other sounds, and is not one which is of great practical value. To meet these difficulties the whisper test has been devised. The procedure is sinular in principle to that used in the watch test, the difference being that the criterion as to whether the subject has heard is in this case the ability to repeat or write numbers which are spoken to him in a whisper. The subject is placed with one ear toward the experimenter, the other ear being plugged, and a position is foxmd as before at which the subject can correctly reproduce numbers which are whispered to him, in eight out of ten cases. In order to insure that the numbers be pronounced with equal loudness each time, the breath should be expelled as fuUy as occurs in a natural expiration, and not more than eight syllables should be pronounced at one time. With these precautions, the whisper test may be made with a fair degree of accuracy, and it may be conveniently used in the schoolroom. c. Acoumeter test. The disadvantage of both the watch and the whisper tests is that they are subject to disturbance by other sounds. Hence, a comparison TESTS 153 between the results with different persons is rendered somewhat uncertain and difficult. These difficulties are overcome by forms of audiometer which produce sounds of regularly varying loudness close to the ear. The instrument which is chosen for this test is the Lehmann acoumeter. This instrument consists essen- tially of a pair of forceps, the height of which can be varied by small and measurable amounts, so as to enable a metal ball to be dropped upon a surface of glass or cardboard from various heights. Directions. Vary the height of the forceps in the acoumeter and determine the threshold in the same manner as in the watch test. Record this as the measure of auditory acuity. Test similarly the other ear. Results of the experiment. The detailed results of the three auditory tests are shown in numerical form in Table XVIII, and in graphic form in Chart VI. The experience of previous classes indicates that when this experiment is made in the usual fashion, that is, when the tests are all made by different individuals, the re- sults are so varied that little or no reliance can be placed on them. This was shown by the unsatisfactori- ness of the checks which are mentioned below. It is evident from this experience that this test, while appar- ently a simple one to make, is one of the most difficult to give accurately. In order to insure greater uniform- ity in the results all the subjects were tested by the watch and the whisper method by one person, and the 1B& EXPERIMENTAL EDUCATION TABLE XVIII. SCORES MADE BY A GROUP OF TWENTY IN THE AUDITORY TEST Acoumeter test Waieh lest Whisj er test Individual R. ear (milli- meters) L. ear (milli- meters) R. ear (meters) L. ear (meters) R. ear (meters) L. ear (meters) A 6. 5. 0. 0. Ai .88 B a. 4. 8.24 1.42 14.42 14.42 C 5. 8. .90 .70 4. 8.0 D 8.5 8.6 1.08 1.14 12.0 8. E 9. a 2.6 0.27 .80 20,22 20.22 F 8. 8. 0.80 10.00 12.00 14.42 G 8. 8. 10.42 10.42 UAt 14.42 H 2.6 2.6 8. 4. 0. 10. I i.5 2.5 2. 2.6 2.26 2.76 1.06 4,80 8,18 1.72 8.40 8. 9.23 11.91 4,20 0.23 J 11.91 K 8.80 L 2. 2,6 4.23 4.23 9.23 0.23 M 2. 2, 6. 4.22 14.12 U.li N 2. 1.6 .0.S 1.88 9,23 10.78 1.75 2.6 1.28 .08 11,6 8.81 P 1.75 2. 4, 2. 7.21 4.9 Q 1.76 1.75 4. 4.22 14.42 UAt R 1.6 8.5 8.11 .00 2.22 8.6 S 1.5 1.6 2. 1.75 .68 2.70 .5 8.8 11.87 16. .08 T 14.42 acoumeter test was given to all by another experi- menter. Even with this precaution the results from five of the subjects, B, E, F, G, and R, were unreliable on account of some limitation or defect in the proce- dure. The data from Subject S are not to be used in judging the reliability of the results because of the loss of one ear-drum. The greatest difBculty encountered in giving this TESTS 157 Distance M 20 m rn n 1 FT" pn rn rr T r 1 1 1 1 1 1 ' ( 1 _ 1 ' ] 1 i __ i [ '~~ ■ ■ 1 J 1 '" -^ . ■ 1 1 o — 1 ^ _ _ g ■ — 1 p 1 1 _ ~ 1 ~ — 1 _ 1 , — I , 1 1 1 ; 1 1 i ' _ ( eriment will serve as examples of tests of progressive difficulty which may be used to measure the degree of mental maturity of an individual. The next step in the devel- opment of maturity tests on a systematic basis is to devise and apply many such single tests to children, in order that norms may be established. By this means a group of standardized tests may be placed at the serv- ice of the teacher or custodian of children, by which ht may analyze the child's mental development quahta- tively and quantitatively. The first test deals with the relatively simple mental process of immediate rote memory, or memory span. Immediate memory develops rapidly with increasing age up to about fifteen years of age, and may be devel- TESTS 163 oped further by practice. This, then, is a suitable sub- ject of a maturity test. Material and method. Various kinds of subject- matter have been used in tests of immediate memory. The most convenient material in a number of ways consists of one-place numbers, and they have accord- ingly been chosen. Numbers have the advantages of being familiar, and at the same time of being relatively free from associations. This insures the use of rote instead of logical memory. Two series of numbers are given below, so that each subject may be given a list with which he is unfamiliar. In making up the lists, any sequence which would be likely to aid the memory of the subject should be avoided. Such would be, for example, regularly as- cending or descending sequences, the immediate repe- tition of the same number, the repetition of the same sequences in successive lists, etc. Ust a: (2) 3, 8 (6) 4, 2, 7, 5, 1, 8 (3) 9, 4, 7 (7) 3, 9, 2, 6, 7, 5, 8 (4) 2, 5, 1, 9 (8) 9, 5, 4, 8, 1, 7, 3, 6 (5) 6, 8, 3, 5, 7 (9) 2, 7, 1, 6, 9, 4, 3, 8, 5 List b: (2) 6, 1 (6) 7, 3, 6, 8, 4, 9 (3) 4, 7, 2 (7) 2, 5, 4, 3, 8, 1, 7 (4) 3, 8, 5, 7 (8) 5, 1, 7, 3, 4, 6, 2, 9 (5) 4, 2, 9, 1, 5 (9) 7, 1, 3, 6, 2, 9, 8, 5, 4 The series are to be read aloud by the experimenter in order, giving the subject opportunity to reproduce each series orally. The numbers should be read in an even voice, without rhythmic grouping at about one second intervals. 164 EXPERIMENTAL EDUCATION b. Reconstruction of sentences Problem. A second test of the same general nature as the preceding, but which involves a more complex sort of mental process, is the subject of this experi- ment. This test consists in rearranging the words of sentences which have been put into chance order. Variation in difficulty has been secured by choosing sentences of the same general character, but of in- creasing length. The sentences used in this experiment have been roughly standardized in difficulty by being given to a class of adults. Material and method. The subject should not try to reconstruct the sentences until the experiment is undertaken. The measure of efficiency is the time required to reconstruct the sentences. Let the experi- menter, with stop-watch in hand, present one sentence at a time to the subject visually, beginning with the shortest and advancing each time to the next longer. The time to be measured is from the presentation of the sentence until it has been correctly written. The answer is to be graded as correct if all the words are included in a grammatical and logical sentence. Minor variations from the key are to be allowed. Two series of sentences are furnished, so that each subject may have a different one. The two series are approximately equal in difficulty. They are printed in the Appendix, which see. Treatment of results. The individual reports should TESTS 165 contain tables showing the series of numbers correctly reproduced and the time required to rearrange each sentence. Any variations from a correspondence be- tween the relative length of a sentence and the time required to construct it should be explained on the basis of introspection. The general report should generalize the data from the individual subjects, and discuss individual diflfer- ences. Results of the experiment. These tests when given to adults are, of course, not tests of maturity. The purpose of including them in a course of this sort, be- sides familiarizing the student with certain typical tests and methods, is to indicate some of the facts which have to be considered in interpreting the results of such tests. The two factors which are to be studied in the results to be presented are individual difference, and the chance misadaptation or favorable adaptation of a subject to an individual test. The immediate memory span for numbers in a group of fifteen individuals varied from 6 to 9. The scores were distributed as follows: — Frequency Number 2 6 3 7 7 8 3 9 The mode of this group is 8. Because of the varia- tion which is here evident among mature individuals. 166 EXPEEIMENTAL EDUCATION it is manifestly a mistake to lay great stress upon the differences of one or two from a norm in a single test such as this. The combined result of a group of tests is of more significance than a single test. Individual differences are also illustrated in the re- sults of the tests in the reconstruction of sentences. These results also show that in this test the time re- quired to reconstruct the sentences depends a good deal on the chance that one begins with the right or the wrong words, or starts the construction of the sentence in one way or another. Table XIX shows the scores of . two groups, in Series I and II respectively. It will be seen upon inspection of the table that there is fairly regular increase in the time required to recon- struct the sentences as they grow progressively longer, but that there are in most cases exceptions to this rule of regular increase, and in some cases the exception is notable. An illustration of a notable exception is to be found in the score of Subject F. Such cases as this show that it would be erroneous to assume — in adults at least — that a sudden and large increase in the time required indicates a limit of ability, or the demarcation between two stages or methods of procedure, for in the last two sentences this subject's score drops consider- ably below the average. It may be that in the case of children or of adults, with still longer and more com- plex sentences, a fairly definite point would be reached which would mark the limit of ability to reconstruct without the expenditure of largely increased time; but. TESTS 167 TABLE XIX. SCORES OF FIFTEEN INDIVIDUALS IN THE RECONSTRUCTION OF SENTENCES Series I Sentence No. Number of words 1 6 7 S 8 4 9 5 10 6 IS 7 13 8 16 Subject — 7 20 30 9 11 13 16 20 10 32 32 14 18 20 25 25 15 19 40 17 23 30 23 22 17 57 32 27 34 27 45 50 21 32 75 19 45 160 51 38 28 56 60 68 117 235 79 41 32 75 40 47 194 47 90 73 36 B 68 C 50 D 52 E 150 p 77 G iiS H 74 Average of middle two individuals 14.5 22.5 22.5 33 41.5 64 60 72.5 Tentative standard 14 22 29 33 47 56 66 Series II Sentence No. Number of vxyrds 1 6 2 7 3 8 i 9 35 16 85 32 27 16 41 5 10 6 12 7 u 8 IS Subject — I 16 12 20 15 14 10 21 17 12 17 14 15 11 19 30 12 20' 17 16 13 36 25 22 24 30 32 16 43 40 31 43 35 37 24 92 37 70 25 60 35 38 181 130 60 J 49 K 45 L 44 M 58 N 36 67 Score of middle individual 15 15 17 32 30 35 60 49 Tentative standard 16 15 20 38.5 51.2 54.25 64.8 168 EXPERIMENTAL EDUCATION with the possible exception of Subject E, no such point was reached in this group, and the probabiHty is against such an interpretation of E's scores. Ejrtension of the experiment. There are several obvious modes of extending this experiment. By fur- ther experiments with adults the standardization of these sentences can be perfected. This might involve modifying some of the sentences, or substituting others for them. Still longer and more complex sentences could be used with adults in the attempt to reach a breaking point. Finally, as in other tests, there is wide opportunity of the standardization of this test with children. TESTS 169 QUESTIONS AND TOPICS FOR DISCUSSION 1. Are any devices used to keep the series of numbers in mind? 2. Discuss the suitability of this method as a test of mental matu- rity. What function is involved? 3. What methods are used to solve the problem of reconstruct- ing the sentences? Are different methods used? Are different methods used with sentences of different length or degree of difficulty? 4. Does your knowledge of child psychology lead you to think that children of different ages might use different methods? 5. What type of mental activity is tested in this experiment? 6. Is the effect of increasing mental maturity to introduce dif- ferent mental processes into the performance of such tasks as these? If not what change does take place? 7. Examine the reports of tests to discover in what kinds of mental process the greatest change with mental maturity ap- pears. ■ 8. Which kind of test, one which shows large, or one which shows small, progress with age, is better suited to the mental ex- amination of children? REFERENCES Squire, Carrie Ransom. "Graded Mental Tests"; in Journal of Educational Psychohgy (1912), pp. 363-80, 430^3, 493-606. Terman, L. M., and Childs, H. G. "A Tentative Revision and Extension of the Binet-Simon Measuring Scale of Intelligence"; in Journal of Educational Psychology (1912), pp. 61, 133, 198, 277. Terman, L. M. The Measurement of Intelligence. Houghton Mifflin Co. (1916). Town, Clara H. The Binet Scale. Translation. Courier Press, Lincoln, 111. Whipple, G. M. Manual of Mental and Physical Tests, pp. 441- 511. Yerkes, R. M., and Bridges, J. W. A Point Scale for Measuring Mental Ability. (Baltimore, 1915.) 170 EXPERIMENTAL EDUCATION Experiment No. 16 correlation between tests Problem. The measurement of correlation is impor- tant in the study of a variety of problems. In educa- tional psychology correlation refers to the extent to which the presence of a certain degree of one mental trait implies the presence of a second trait in a corre- sponding degree. Put in another way, the degree of correlation between two traits may be measured by the degree of correspondence between the two series or orders when the individuals of a group are ranked according to their eflSciency in each of the two proc- esses. Some of the various methods of calculating degrees of correlation are therefore based upon the arrange- ment of a group of individuals in ranks. For example, suppose we wish to determine the correlation between height and weight in a group of persons. After each in- dividual has been weighed and measured the individ- uals are given a ranking according to each of the two characteristics separately. If the same person stands at the head of the two lists, — that is, if the tallest per- son is also the heaviest, — and if each person occupies the same position in the two rankings all the way down, there is complete positive correlation. On the other hand, if the individual who is at the top of one ranking is at the bottom of the second, and if this relation is TESTS 171 maintained throughout, there is complete negative correlation. If such a relation obtains between the two rankings as would be ascribed only to the operation of chance, there is no correlation. More exact methods take account not simply of the ranJc or position of an individual, but also of the amount of his deviation from a central tendency. For the study of the correlation between mental traits various methods have been used. A brief ac- count of the more important methods is given in the discussion of the results of this experiment. For fuller discussion the reader is referred to the convenient ac- count which is given by Whipple in his Manual of Mental and Physical Tests. Whipple appends a bibU- ography which may be used as a guide to further study. Examples of the extensive use of the correlation method in psychology are to be found in the investiga- tions of Burt and Simpson. These and other investigations have made it clear that there is a high degree of correlation between some mental traits, and a low degree between others. The purpose of this experiment is to measure the correla- tion between mental traits which are representative of various groups of mental abilities. Three traits or forms of ability will be measured, one of them being tested twice, and the four correlations between the various pairs will be found. Material and method. The abilities which are to be measured are (1) rate of tapping, as an illustration of 172 EXPERIMENTAL EDUCATION motor ability; (2) pitch discrimination, as an illustra- tion of sensory discrimination; and (3) the opposites test, as an illustration of the more complex mental processes. (1) The tapping test is a simple one to perform. By means of a stop-watch and an apparatus for recording the number of taps the rapidity of tapping with the hand is measured. (See Figure 7.) The number of taps CmSns \ Fio. 7. DIAGRAM OF THE CONNECTIONS FOR THE TAPPING APPARATUS is recorded by an electric counting device. A stylus and brass plate are in circuit with an electric battery or other circuit and with an electric clock. Each time the circuit is closed by bringing the stylus in contact with the brass plate, the hand of the clock moves for- ward one point. The procedure then is to tap continu- ously for five seconds, as rapidly as possible, and then read off the record on the clock. The clock should be set at zero before each trial by turning the hand for- ward. In tapping, the forearm should be allowed to TESTS 173 rest on the table, and the movement should be made with the wrist. Three trials of five seconds duration each should be made, and the average of the three taken. , (2) The test in pitch discrimination consists in the determination of the amount of difference between two tones which must exist in order that one may be distinguished as higher than the other. A convenient and accurate form of apparatus to use in measuring pitch discrimination is the tuning fork. In the present experiment a standard tuning fork of 435 vibrations is used, and a number of other forks differing from it by varying numbers of vibrations for making the com- parisons. The same general method of procedure is to be used as in the test of auditory acuity : that is, the compari- son should begin with tones which can be easily dis- tinguished and proceed gradually to the smaller inter- vals, until a point is reached at which eight out of ten judgments are correct. The interval between the two tones as found thus may be recorded as the descending threshold. This threshold is sufficiently reliable. Several features of procedure should be mentioned. The fork which is to be compared with the standard should first be put in position with the standard fork on the sounding box. Then, after a " ready " signal, one of the forks should be struck, then damped, and then the second struck and damped. While one fork is sounding the other one should be damped to pre- 174 EXPERIMENTAL EDUCATION vent sympathetic vibration. The duration of each tone and of the interval between them should be uni- formly about two seconds. Care should be taken to avoid any regularity in striking either the standard or the comparison tone first. Occasionally the same tone should be struck twice in succession as a check. (3) The procedure in the opposites test is simple. A list of words is furnished and the problem is to supply for each one a word which has the opposite meaning. The task is to supply the opposite word within a time limit of ten seconds for each word. The response words should be written, and the time may be extended until the writing of a word is completed if it has been begun within the ten seconds. In order that the same test may be used by subject and experimenter, the sub- ject should take the printed list of words and cover it with a sheet of paper. Each word may be uncovered by sliding a card so that the opening comes over it at a signal given by the experimenter at the interval men- tioned. The subject should study the word until the next signal is given or until the opposite is found. The score consists in the total number of opposites cor- rectly given. The response, to be correct, need not be the exact word given in the response list, but should be a synonym of it. The lists of stimulus words and specimen opposites are given in the Appendix, which see. Treatment of results. Since the chief results in this experiment appear from a comparison of the results TESTS 175 from the various individuals of the group, each mem- ber of the class should calculate one correlation coeflS- cient. The members of the class may be numbered. Number 1 should calculate the first correlation men- tioned below, number 2 the second, and so on. Num- ber 5 should begin at the first, and so on. The correla- tions to be calculated are as follows: — 1 . Between rate of tapping and pitch discrimination. 2. Between rate of tapping and first opposites test. 3. Between pitch discrimination and first opposites test. 4. Between first and second opposites tests. For convenience in arranging the list of subjects, in order to calculate the correlations, they should be des- ignated by their number. Each person should there- fore indicate his number on his report. Each individual report should include, beside the usual preliminary statement of the problem and method, the full account of the results of each test, including introspections and description of the diffi- culties which were encountered. The individual report should also include the full calculation of the correla- tion coefficient, and not merely the coefficient itself. The general report should compare the degrees of correlation among the results of the various tests, and should enter upon a discussion of the significance of the differences found. Reference should be made to at- tempted explanations of correlation and differences in correlation among mental processes, such as Krueger 176 EXPERIMENTAL EDUCATION and Speaxman's theory of a central intelligence or of a hierarchy of intelligences. Results of the experiment. The scores in the three tests used in this experiment are given in Table XX, TABLE XX. SCORES OF NINETEEN INDIVIDUALS IN THE TESTS Individual RaUof Pitch dis- Opposites tapping criminaiion test I 86 8 m 38 i 20J 38| 1 m 40 17 26 40| 2 271 41i 2 m 41f 8 16 42 3 ISJ 43 17 22 431 i 20^ 45 i 18J 46 1 15 46 i 23J 47 2 17f 47i i 24 47i i 24 49- i 26 53 2 19§ •• 8 26 Opposites test J I 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17, 18, 19. 10 20 13 11 6 10 17i 10 11 10 16 18i 18 13 with the subjects arranged in their order in the tapping test. Charts "\7I and VIII and Table XXI illustrate methods of displaying and calculating the correlation between the first and second opposites tests. These results are presented more for the purpose of illustrat- ing some of the methods of examining correlation and II 20 16 Score 20 >16 10 TESTS A 177 ~~ ~~ ~~ ~ ~ ■" ~ ~ ~ ~ • ~ I 15 20 25 B — J y — T r — — II IndiTldual ABCDEFGHIJKLMN Chaet VII. GRAPHIC REPRESENTATION OF THE CORRELATION BETWEEN THE FIRST AND SECOND OPPOSITES TESTS some of the requirements of valid procedure than as a basis for much discussion of the facts of correlation themselves. Table XXI illustrates a form of procedure which is necessary, in many cases, to obtain a reliable calcula- tion of correlation, that is, the determination first of the reliability of the measures secured in each test by 178 EXPERIMENTAL EDUCATION itself. This is seemed by finding the cocrelation be- tween two peif onnances in the same test, using, where the nature of the test demands it, different sobject- TABLE YXT. CORBEIATIOX BETWEEN" FIBST AXD SECOND OPPOSITES TESTS Store inl Sean in II icons in Ifrom az^age 9 dif.af frorfi in 11 from aterage x^ y' rg 6.. 7.. 8.. 9.. 10.. 11.. 12.. 13.. 14.. Aveiage. 15 15.3 16 17.5 17.5 17.5 18.5 19.5 20.5 20.5 20.5 *2 23.5 H 10 10 6 10 11 18.5 11 13 10 13 20 17.5 16 18 -4 -3. —3 -1. —3 -3 —3 -1.5 f -2 -1.5 ^5 - .5 -2 + ^ i +1.5 —3 i -rl.5 1 +1.5 1 -^7 ' +3 —4 • +4.5 —3 16 I 12.25 2.25 I 2.25 2.25 .25 .23 2.25 2.25 ! 2.25 I 9 \ 20.25 io 9 9 49 9 4 30.25 4 9 49 12 +10.5 +21 4.5 + 3.0 - 8.25 + 1 - 4.5 +10.5 20.25 +13.5 9 +13.5 25 :+25 19 13 105.5 226.5 101. 1x- y V2l>- 2»" VU6.3X2263 154.6 snm of tiie products of x and y \ ' square root of (the sam of 2^ X ihe sum of if) J matter ia the two pecf onnances. If this correlation is not fairly high — above .60 — the degree of correla- tion between this test and others is of little significance, since the scores are not accurate measures of the ability TESTS 179 in question. A formula has been developed by Spear- man to correct a coefficient of correlation when it is Positive cortelatioD ZTegatiTe conelatkm < 1 M ' 1 1 1 1 ; I ' ' ! : ' ' \ y, (■-« 1 i yf t ' ' X T : • : : ' ■ X: ; , ' ' ' y r^ X A £ ' ' 1 ; y^' u ■r. \ ' 'r /. ■ \\ ' ! ■ 2 ' i y ' ' X^ ' M i yT ■ f 1 ! i in: : 1 1 ' .ABcending — »■ ■ 3r /~ /-/ — z.:^__ trn / / -r,§^ ^ / B« z; ^ _- -*^g ^.:n), pp. 96 f. 186 EXPERIMENTAL EDUCATION Krueger, F., and Spearman, C, "Die Correlation Zwischen ver- scheidenen geistigen Leistungsftihigkeiten"; in Zeilschrift fur Psy- cliologie, etc. (1907), vol. 44. pp. 50-114. Ries, Georg. "Beitrage sur Methodik der Intelligenz-prtlfung"; in Zeilschrift fur Psychologic, etc. (1910), vol. 56. pp. 21-343. Simpson, B. R. Corrdationa of Mental AbUMes. (Columbia Uni- versity Press.) Spearman, C. "Greneral Intelligence Objectively Determined and Measured"; in American Journal of Psychology (1904). vol. 16, pp. 201-9S. Stern, W. Psychological McUiods of Testing InUUigence. Tr. by G. M. Whipple. (Warwick & York, Baltimore.) Thorndike. E. L. Educational Psychology, vol. m. Thorndike, E. L. An Introduction to the Theory of Mental and Social MeasurerMnts. (New York, 1904.) Whipple, G. M. MantuU of Menial and Physical Tests, vol. 1, chap. III. APPENDIX FIGURES AND TESTS FOR USE WITH THE EXPERIMENTS (189) (190) \ \ / FIGURE FOR EXPERIMENT 3 — TAIT UNICUBSAL LABYRINTH (191) SEEIES OF SYLLABLES FOR EXPERIMENT 5 Series I Series la Series II geb pos wix wap gaj pif ker yed zin zub bik faq ren nal pex yat wur zan ref juv fos biv mip nel lin buv ber taz yot rul fud dik kif zib caz lod mep mub kir dos gom ral rel cug fup Series III Series IV Series V zed paf pom vof kuj loy lud dal feg flp ros hoi mol tiv cag lat kay kas ced tud lira mai soz rin sug mez nop rol gir det dut fon tur taj vid rix pud sim lez hov kaq gim vul eer dak POEM TO BE USED IN EXPERIMENT 6 ENOCH ARDEN Long lines of cliff breaking have left a chasm; And in the chasm are foam and yellow sands; Beyond, red roofs about a narrow wharf In cluster; then a moulder'd church; and higher A long street climbs to one tall-tower'd mill; And high in heaven behind it a gray down With Danish barrows; and a hazelwood, By autumn nutters haunted, flourishes Green in a cuplike hollow of the down. Here on this beach a hundred years ago. Three children of three houses, Annie Lee, The prettiest little damsel in the port. And Philip Ray, the miller's only son. And Enoch Arden, a rough sailor's lad Made orphan by a winter shipwreck, play'd Among the waste and lumber of the shore, Hard coils of cordage, swarthy fishing-nets, Anchors of rusty fluke, and boats up-drawn; And built their castles of dissolving sand To watch them overflow'd, or following up And flying the white breaker, daily left The little footprint daily wash'd away. A narrow cave ran in beneath the cliff; In this the children play'd at keeping house. Enoch was host one day, Philip the next. While Annie still was mistress; but at times Enoch would hold possession for a week: "This is my house and this my little wife." "Mine too," said Philip; "turn and turn about"; When, if they quarrell'd, Enoch stronger-made Was master. Then would Philip, his blue eyes All flooded with the helpless wrath of tears. 194 APPENDIX Shriek out, "I hate you, Enoch," and at this The little wife would weep for company, And pray them not to quarrel for her sake, And say she would be little wife to both. But when the dawn of rosy childhood past, And the new warmth of life's ascending sun Was felt by either, either fixt his heart On that one girl; and Enoch spoke his love, But Philip loved in silence; and the girl Seem'd kinder unto Philip than to him; But she loved Enoch, tho' she knew it not. And would if ask'd deny it. Enoch set A purpose evermore before his eyes. To hoard all savings to the uttermost. To purchase his own boat, and make a home For Annie; and so prosper'd that at last A luckier or a bolder fisherman, A carefuUer in peril, did not breathe For leagues along that breaker-beaten coast Than Enoch. Likewise had he served a year On board a merchantman, and made himself Full sailor; and he thrice had pluck'd a life From the dread sweep of the down-streaming seas. And all men look'd upon him favorably. And ere he touch'd his one-and-twentieth May He purchased his own boat, and made a home For Aimie, neat and nestlike, halfway up The narrow street that clamber'd toward the mill. Then, on a golden autumn eventide. The younger people making holiday. With bag and sack and basket, great and small. Went nutting to the hazels. Philip stay'd — His father lying sick and needing him — An hour behind; but as he climb'd the hill. Just where the prone edge of the wood began To feather toward the hollow, saw the pair, Enoch and Annie, sitting hand-in-hand. APPENDIX 195 His large gray eyes and weather-beaten face All-kindled by a still and sacred fire. That burn'd as on an altar. Philip look'd And in their eyes and faces read his doom; Then, as their faces drew together, groan'd. And slipt aside, and like a wounded life Crept down into the hollows of the wood; There, while the rest were loud in merry-making, Had his dark hour unseen, and rose and past Bearing a lifelong hunger in his heart. So these were wed, and merrily rang the bells. And merrily ran the years, seven happy years, Seven happy years of health and competence, And mutual love and honorable toil. With children, Qrst a daughter. In him woke. With his first babe's first cry, the noble wish To save all earnings to the uttermost. And give his child a better bringing-up Than his had been, or hers; a wish renew'd, When two years after came a boy to be The rosy idol of her solitudes. While Enoch was abroad on wrathful seas, Or often journeying landward; for in truth Enoch's white horse, and Enoch's ocean-spoil In ocean-smelling osier, and his face, Rough-redden'd with a thousand winter gales. Not only to the market-cross were known. But in the leafy lanes behind the down. Far as the portal-warding lion-whelp And peacock yew-tree of the lonely HaU, Whose Friday fare was Enoch's ministering. Then came a change, as all things human change. Ten miles to northward of the narrow port Open'd a larger haven. Thither used Enoch at times to go by land or sea; And once when there, and clambering on a mast In harbor, by mischance he slipt and fell. 196 APPENDIX A limb was broken when they lifted hinti; And while he lay recovering there, his wife Bore him another son, a sickly one. Another hand crept too across his trade Taking her bread and theirs; and on him fell, Altho' a grave and staid God-fearing man. Yet lying thus inactive, doubt and gloom. He seem'd, as in a nightmare of the night. To see his children leading evermore Low miserable lives of hand-to-mouth. And her he loved a beggar. Then he pray'd, "Save them from this, whatever comes to me." And while he pray'd, the master of that ship Enoch had served in, hearing his mischance. Came, for he knew the man and valued him. Reporting of his vessel China-bound, And wanting yet a boatswain. Would he go? There yet were many weeks before she sail'd, Sail'd from this port. Would Enoch have the place? And Enoch all at once assented to it. Rejoicing at that answer to his prayer. So now that shadow of mischance appear'd No graver than as when some little cloud Cuts off the fiery highway of the sun. And isles a light in the offing. Yet the wife — When he was gone — the children — what to do? Then Enoch lay long-pondering on his plans : To sell the boat — and yet he loved her well — How many a rough sea had he weather'd in her! He knew her, as a horseman knows his horse — And yet to sell her — then with what she brought Buy goods and stores — set Annie forth in trade With all that seamen needed or their wives — So might she keep the house while he was gone. Should he not trade himself out yonder? go This voyage more than once? yea, twice or thrice — As oft as needed — last, returning rich. Become the master of a larger craft. APPENDIX 197 With fuller profits lead an easier life. Have all his pretty young ones educated. And pass his days in peace among his own. Thus Enoch in his heart determined all; Then moving homeward came on Annie pale. Nursing the sickly babe, her latest-born. Forward she started with a happy cry. And laid the feeble infant in his arms; Whom Enoch took, and handled all his limbs. Appraised his weight and fondled fatherlike. But had no heart to break his purposes To Annie, till the morrow, when he spoke. Then first since Enoch's golden ring had girt Her finger, Annie fought against his will; Yet not with brawling opposition she. But manifold entreaties, many a tear. Many a sad kiss by day, by night, renew'd — Sure that all evil would come out of it — Besought him, supplicating, if he cared For her or his dear children, not to go. He not for his own self caring, but her. Her and her children, let her plead in vain; So grieving held his will, and bore it thro'. For Enoch parted with his old sea-friend. Bought Annie goods and stores, and set his hand To fit their little streetward sitting-room With shelf and corner for the goods and stores. So all day long till Enoch's last at home, Shaking their pretty cabin, hammer and axe. Auger and saw, while Annie seem'd to hear Her own death-scaffold raising, shrill'd and rang. Till this was ended, and his careful hand, — The space was narrow, — having order'd all Almost as neat and close as Nature packs Her blossom or her seedling, paused; and he, Who needs would work for Annie to the last. Ascending tired, heavily slept till morn. 6 ^ a s^ «.§ i af 13 •" g 0} o o 3 O 13 o CD 0) .c3 -hi is PI O '53 ^^ O o ID !i! 1) bo 133 O O ^ 0) cd en •^ '^ -4-> rd 0) -I-' PI a ^ M o3 o a o o en ■ i-t ■ — ! "T^i if. a S ^ a ■: 0) 1) 0) ID ^ ■u el g, }-* tn 3 5 o3 ^ "o -i d D o3 fj tn .^ .tn _tn "C be •5 "J "5 >-i '-J n3 ^ 'H tn I— I I— I o o o o ^ rd o o tn tn 4) o3 '^ a O d T3 ,d 4) O (-1 •'3 O ^ '5 03 4) -4-} > a> ^ -^ rd ^ tu d "3 T3 d tn o d a. 2 d ^ .a o ft"*- > 'a 03 g bO '^ d "C rd ,o ■" ■- Sf rd .d O " (U t--H ^ tn •?.a (198) (199) 200 APPENDIX No. 2— 11-PoiNT Type, 12 Ems Line No elaborate discussion or criticism of the methods by which the preceding material has been gathered can be at- tempted. We have gone on the assumption that the various studies have been, at least in general, reliable. The interests that have been mentioned are, most of them, well recognized by all observers of children, and, while there may be errors in observing and interpreting the facts here set forth, the general scheme as a whole seems fairly consistent when we reflect upon the great variety of sources from which the material has come. The whole subject of methodology does, however, need a thorough investigation. The object of all this class of studies should be to get ac- curate data as to the chUd's spontaneous expressions and ac- tivities, with definite record as to age, sex, and previous life- APPENDIX 201 history. Every study must be accompanied by a careful state- ment of the conditions under which the material is secured. The real child-psychologists are endeavoring more than ever before to devise tests that will eliminate as nearly as possible tj ii 5=3 trt ;g a . o i .s o :s « K a B M §^^•3 2 o § a te ffi s M S S h S K H, O. a ^^-S S 3 ^ .-s (J O ^ o ■a a - y? o ni ■s-g (4-< in - 3 g ■S '5 5 -2 "S g >, & +j -rt »s ' ^ ,.? ?J 13 -^ Cd n^ S -4-) en rt S ;=< , fl JS ° " • •2 j3 .2 .3 33 _, tj bo » 2 I ° M ■rl .S h S " "3 s Si I s a £ -S 5 u o a o § !3 - g I rt 3^0 ' .s ^ c Q rt Ci ■3 .5 3 «|.s S m K ' P u O O [A O , +j ia M Pi w J5 '-S -^ -S <- *^ ts a — QJ E x{ ^ TO CO 4) - ^ a .s u ,a a c9 ■ •" o t; S ■« " t; o ° " 2 § i^ £ o a a ■t|2g ■3 iJ I 3 S.S I S^ " -a >> " .55 >> ff ■■"2^3 •^ 13 K -a m to S a 0) -a 5- I "^ 1! ■« « I -g .^ 5 -s o Eio ,« tj rt rf M o ^^ m eiJ S CO +j " 5 -a o " si i a o • ij_S Hj • • • • A • '^^'^ i A A* mm to m m ^ cii^ • •• d"§ HI -^ H g % • • • • • • • i| (212) APPENDIX 213 MATERIAL FOR USE WITH EXPERIMENT 15 PART B, RECONSTRUCTION OF SENTENCES Series I No. of words „ an boy the apple greedy ate dog his faithful 7 the slowly followed master wolves outran the 8 swift pursuing horse the soon footsteps of they 9 soon the heard afterward sound many 10 quickly the as cat flew siezed locust the up it No. of words clearing on as the side they fugitive other the entered the disappeared 12 14 16 and his closed secretary the himself office door the seated of desk the at watch the train at would his late traveler twenty the knew minutes and be glanced that 214 APPENDIX Series II No. of words „ hungry bone the gnawed dog his ran the ditch 7 horse the into frightened very prize the 8 hard Mary and won studied some said the 9 steam boy saw he that yachts 10 and slowly the channel ship thru the narrow long sailed No. of words every the approached herd cautiously it side hunters the and on surrounded 12 the fence-tops all had snow when 14 day the drifts up the to were fallen shed the further their departed tools 16 in workmen disturbance homes placed the without and to their APPENDIX 215 EXPERIMENT 15 — KEY TO SENTENCE ARRANGEMENT Time Series I {including loriliny) The greedy boy ate an apple U The faithful dog followed his master slowly 22 The swift horse soon outran tlie pursuing wolves ... 29 Soon afterward they heard the sound of many foot- steps 33 As the locust flew up the cat quickly seized it ... . 39 ' As they entered the clearing the fugitive disap- peared on the other side 47 The secretary closed the door of the ofEce and seated himself at his desk 56 The traveler glanced at his watch and knew that the train would be twenty minutes late 66 Series II The hungry dog gnawed his bone 16 The frightened horse ran into the ditch 15 Mary studied very hard and won the prize 20 The boy said that he saw some steam yachts 30 The ship sailed thru the long and narrow channel . 38 . 5 The hunters approached the herd and cautiously surrounded it on every side 51.2 When the snow had fallen all day the drifts were up to the fence-tops 54 . 25 The workmen placed their tools in the shed and departed to their homes without further dis- turbance 64 . 8 ' Estimated. 216 APPENDIX STIMULI FOR USE WITH EXPERIMENT 16 Opposites Test Stimulus List No. 1 1. intelligent 2. however 3. enthusiastic . . . 4. traitor 5. intricate 6. sublime 7. petty 8. languor 9. disdain 10. thorough 11. vacillating .... 12. fastidious 13. important .... 14. spendthrift . . . . 15. motion 16. dextrous 17. serious 18. gentle 19. unless 20. although 21. prohibit 22. uncouth 23. conceal 24. precise 25. rigid 26. suave 27. proficient 28. belief 29. cruel 30. result Stimulus List No. 2 1. to respect 2. to hold 3. exciting 4. simple 5. deceitful 6. permanent .... 7. to degrade 8. level 9. suspicious 10. pride 11. despondent . . . . 12. venturesome . . . 13. silly 14. busy 15. preserve 16. abet 17. abeyance 18. abnegation .... 19. absolve 20. alternative .... 21. captious 22. equivocal 23. fame 24. hazard 25. hypocrisy 26. imminent 27. inherent 28. pique 29. satisfy 30. terse APPENDIX 217 RESPONSE LIST FOR EXPERIMENT 16 Opposites Test Response List No. 1 1. stupid 2. accordingly (hence) * 3. indifferent 4. patriot 5. simple 6. commonplace 7. noble 8. energy, vigor 9. respect 10. superficial 11. constant 12. ne^igent, slip-shod 13. insignificant 14. miser 15. rest 16. clumsy 17. frivolous 18. rough 19. if 20. because 21. permit 22. cultured, polished 23. reveal 24. vague, careless 25. pliable 26. brusque 27. unskilled 28. doubt 29. sympathetic 30. cause Response List No. S 1. despise 2. to release 3. soothing 4. complex 5. frank, truthful 6. temporary 7. to ennoble 8. tilted 9. trustful 10. humility 11. cheerful 12. cautious 13. sensible 14. idle 15. destroy 16. frustrate, impede, hinder 17. enforcement, operation 18. claim, demand 19. condemn 20. compulsion, necessity 21. fair 22. plain, unambiguous 23. oblivion 24. certainty 25. sincerity 26. improbable, contingent 27. incidental, unconnected 28. complacency 29. disappoint 30. diffuse Synonyms may be allowed. Do not use the piefix un. INDEX Aeoumeter, 154. Age differences, 23, 30. Analysis, 32/. Analytical scale, 83, 89, 91, 92. Astigmatism, 141/. Audiometer, 152. Ayer, F. C, 31. Ayres, L. P., 82, 94. Ayres scale, 83, 89, 90, 91, 92. Bach, T., 151, 161. Bagley, W. C, 41, 185. Bair, J. H., 13, 24. Bergstrom, J. E., 49. Binet-Simon measuring scale, 169. Book, W. F., 13, 24. Bridges, J. W., 169. Brown, W., 180, 185. Bryan, 6, 13, 20, 24. Burt, C, 171, 184, 185. Childs, H. G., 169. Clement, J. A., 185. Colvin, S. S., 50, 63. Coordination, motor, 73/. Correlation, 60, 115, 119jf., 170/. Courtis, S. A., 138. Cowling, D. J., 27, 31. Dearborn, W. F., 24, 107, 109, 111, 183, 185. Dewey, J., 40, 139. Dodge, R., 95, 107. Drawing, 25/. Ebbinghaus, H., 63, 71. Elliot, 87. Erdmann, 95. Eulenberg, H., 151, 161. Experiment, 5. 7, 8, 10, 158/. Experiment, figures for drawing, 189/. Experimentation, method of, 4/.; rules for, 9/. Form in handwriting, 82 /. Freeman, F. N., 81, 94, 131, 139. Gilbert, C. A., 185. Grading, reliability in, 87/ Gray, C. T., 92, 93, 94. Grouped objects, arrangement of, 208/. Grouping in number apprehen- sion, 133/ Hand tracer, 73. Handwriting, analysis of, 72/ Handwriting scales, 83, note. Harter, 6, 13, 20, 24. Heck, W. H., 50, 185. Heterophoria, 141/ Huey, E. B., 95, 107, 109. Hyperopia, 41/ Identical elements in transfer, 147/. Individual differences, 20/., 38/., 53, 00, 67/., 75, 114, 120/, 137/, 147/., 156/, 165/., 176. Instruction in learning, 17/., 28, 32. Judd, C. H., 27, 31, 81, 108, 109, 116, 119, 131, 139. Kelley, T. L., 90. Kinetoscopic method, 108. King, Irving, 198. Krueger, F., 175, 186. Labyrinth, Tait, 36. Lange, R., 63. 220 INDEX Lay, W., 51, 63. Learning, experiments in, 2 /., perceptual, 25 jf. problem solving, 32. sensori-motor, 13^. types of, 12. Lehman acoumeter, 155. Lindley, E. H., 36, 40. Maddox rods, 140. Manuel, H. T., 90, 94. McAllister, C. N., 109. McClellan, J. A., 139. McCallie vision test, 143. Memorizing, of sense material, 64^. part method of, 65^. permanence, 54/. recall during memorizing, 55 /. rote, 51/. whole method of, 65 jf. Memory, immediate, 162/. Mental tests, 3/. Messmer, 114. Meumann, E., 51, 63, 71, 114. Mirror for observing eye move- ments, 95. Mirror tracing apparatus, 14/., 41/. Mnemonic devices, 53/. Miinsterberg, H., 138. Myopia, 141/ Nonsense syllables, 192. Observation as a scientific method, if. Opposites test, 174 /.; material for, 216/ Pearson, 120, 180, 183. Pintner, R., 90. Pitch discrimination, 173/ Poem for memorizing, 193/ Psychology, educational, 1/ Puzzle box, 34/ Reading, efiSciency in, 117/ eye movements, 105/ fusion of elements in, 113/ Reconstruction of sentences, 164/.; material for, 213/ Ries, G., 184, 186. Ruger, H. A., 24, 40. Simpson, B. R., 171, 184, 186. Spearman, C, 176, 179, 180, 186. Squire, Carrie Ramson, 169. Starch, D., 24, 87, 90, 131. Steele, W. M., 109. Stenopaic lens (Stevens), 146. Stern, W., 186. Swift, E. J., 13, 20, 24. Tachistoscope, 111. Tachistoscopic method, 110, 133. Tachistoscopic stimuli, 212. Tait labyrinth, 191. Tapping board, 172. Tapping test, 172/ Terman, L. M., 169. Texts for study of eye-move- ments in reading, 198/ Thorndike, E. L., 40, 50, 82, 83, 93, 94, 131, 186. Thorndike scale, 83. Town, Clara H., 169. Transfer for training, 41 / Trial frames, 143. Trial lenses, 143. Vision test cards, 143. Waldo, K. D., 131. Wallin, J. E. W., 7. Watt, H. J., 71. Weis, A. P., 90. Whipple, G. M., 8, 63, 114, 115, 116, 120, 151, 161, 169, 171, 180, 186. Yerkes, R. M., 169.