A COMPARATIVE STUDY OF DIRECTED AND UNDIRECTED TEACHING BY FRANCIS SHREVE. PH. D. GEORGE PEABODY COLLEGE FOR TEACHERS CONTRIBUTION TO EDUCATION NUMBER SEVEN PUBLISHED UNDER THE DIRECTION OF GEORGE PEABODY COLLEGE FOR TEACHERS NASHVILLE, TENN. 1922 A COMPARATIVE STUDY OF DIRECTED AND UNDIRECTED TEACHING BY FRANCIS SHRE\ E, PH. D. GEORGE PEABODY COLLEGE FOR TEACHERS CONTRIBUTION TO EDUCATION NUMBER SEVEN PUBLISHED UNDER THE DIRECTION OF GEORGE PEABODY COLLEGE FOR TEACHERS NASHVILLE. TENN. 1922 Copyright, 1922, by Francis Siireve McQuiDDY Printing Co. Nashville CONTENTS Chapters Page I. Introduction 5 Review of Literature 5 Statement of the Problem 25 II. The General Plan of the Study 27 Plan I: Two Teachers 27 Plan II: One Teacher 27 A Specific Plan for Each Experiment 28 Variable Factors 29 III. The Experiments 32 Experiment 1 : 32 Experiment 2 39 Experiment 3 48 Experiment 4 52 Experiment 5 57 Experiment 6 62 Experiment 7 67 Experiment 8 72 IV. Summary and Conclusions 75 Summary 75 Conclusions 78 Bibliography 80 ACKNOWLEDGMENTS The writer takes pleasure in acknowledging his indebted- ness to the following superintendents, principals, and teach- ers whose hearty cooperation made this study possible : Supt. 0. G. Wilson, P'airmont, W. Va. Principal G. H. Colebank, Fairmont, W. Va. H. Y. Clark, Fairmont, W. Va. Miss Elizabeth Koletka, Fairmont, W. Va. Miss Frances Frost, Fairmont, W. Va. Supt. Edgar B. Simms, West Union, W. Va. Principal Howard H. Jones, West Union, W. Va. Miss Orpha Ashburn, West Union, W. Va. Mrs. Lyda M. Strickling, West Union, W. Va. Miss Bessie I. Nutter, West Union, W. Va. Supt. Goff D. Ramsey, Pennsboro, W. Va. Ira Taylor, Pennsboro, W. Va. Mrs. Mary P. Woodell, Pennsboro, W. Va. C. N. Baldwin, Pennsboro, W. Va. Principal M. H. Cole, Fairmont. W. Va, Miss Mary Frazier, Fairmont, W. Va. Miss Lulu M. Jones, Fairmont, W. Va. Miss Launa E. Mason, Fairmont, W. Va. Miss Helen M. Fleming, Fairmont, W. Va. Miss Evelyn Prickett, Fairmont, W. Va. Miss Willard Clayton, Fairmont, W. Va. Supt. John C. Shreve, Littleton, W. Va. Miss Anna C. O'Dea, Littleton, W. Va. Principal Otis H. Milam, Barrackville, W. Va. Miss Beatrice Tennant, Barrackville, W. Va. The writer's greatest obligation is to Dr. Charles A. Mc- Murry, Dr. Joseph Peterson, and Dr. Shelton Phelps, for helpful criticism and wise counsel. He is greatly indelDted, also, to his wife, Elma Cobb Shreve, for invaluable aid and encouragement during the two years required for the com- pletion of the investigation. A COMPARATIVE STUDY OF DIRECTED AND UNDIRECTED TEACHING CHAPTER I INTRODUCTION This study undertakes to evaluate one phase of the su- pervision of instruction on the basis of quantitative data. This method of measuring educational progress is rapidly displacing the earlier method of personal opinion. The last decade has witnessed a rapid development of standard scales and measurements, w^hich has made possible the study of educational problems on a more scientific basis. The literature in this field is already abundant and is rap- idly accumulating. The abundance of material, however, imposes the necessity of confining the review which is to follow to the literature bearing most directly on the prob- lem under investigation. I. Review of Literature For convenience of treatment, the books and articles bearing on the present problem may be classified roughly into three groups: (1) those dealing with the standard tests and their uses; (2) experimental studies in the field of method; and (3) those dealing with directed learning or supervised study. Group I: Standard Tests and Measurements The work in this field has been summarized in readable form in a number of recent books — namely : Starch, Daniel — Educational Measurements ; Monroe, W. S., et al. — Educational Tests and Measure- ments ; Chapman, J. C, and Rush, G. P. — Measurement of Class- room Products ; Monroe, W. S. — Measuring the Results of Teaching; Wilson, G. M., and Hoke, K. J. — How to Measure. These books are well known to all scientific students of educational problems. A detailed treatment of this type of literature is unnecessary, therefore, for our present pur- pose. One general tendency in this field, however, should be noted. The first scales were designed to measure gen- eral merit only, while many of the later scales have been 6 A Comparative Study of designed to measure particular characteristics. The former may be said to be nonanalytic, whereas the latter are an- alytic or diagnostic. This tendency is clearly exhibited, for example, in the development of handwriting scales. Thorndike's Handwriting Scale, appearing in 1910, is a scale for measuring the general merit of handwriting; but Freeman's Scale, appearing in 1915, is analytic. It at- tempts to measure the essential characteristics of hand- writing — namely, uniformity of slant, uniformity of align- ment, letter formation, spacing, and quahty of line. The same tendency is seen in composition scales. The Hillegas Scale measures general merit, while the Harvard-Newton Scale differentiates the four typical forms of composi- tion : narration, description, exposition, and argumentation. Moreover, the defects and merits of each sample are noted on the scale. Similarly, in other subjects diagnostic scales are being devised. This tendency is noteworthy, for it in- dicates that the standard scales are being utilized more and more to diagnose the pupil's difficulties as a basis for plan- ning remedial measures. The magazine articles dealing with the standard scales are too numerous to receive individual attention. They deal particularly with the construction of scales and their uses in measuring the general merits of educational prod- ucts and in diagnosing the pupil's abilities and difficulties in the various school subjects. These articles will be listed in the general bibliography. Group II : Studies in the Field of Method Numerous experimental studies have been reported deal- ing with various phases of method, but only those most closely related to the present investigation will be reviewed in this connection. Several studies have been made to determine the value of direct or formal drill in school subjects. Does drill pay? What type of drill is most effective? These are the ques- tions that have received most attention. Phillips performed an experiment to determine the value of drill in arithmetic. He used as subjects 69 pupils in the sixth, seventh, and eighth grades in one school, divid- ing them into two groups approximately equal in size and initial ability as determined by the scores on the Stone tests for fundamentals and for reasoning. The drill group was given ten minutes' drill daily, subtracted from the class period, for two months, while the non-drill group continued the regular class work. At the close of this period both groups were tested again with the Stone tests. It was Directed and Undirected Teaching 7 found : "That the improvement in the fundamentals of the combined drill groups was 15 per cent greater than the non-drill groups." In every case, with one exception, the drill groups excelled the nondrill groups. But the two groups may have been unequal in the ability to learn, or the teachers conducting the drill may have been superior in teaching ability to the teachers instructing the non-drill groups. The report does not indicate that any at- tempt was made to equalize the learning ability of the two groups of pupils or the teaching ability of the two groups of teachers. Moreover, the difference of attainment of the two groups is too small to have much significance. (Jour, of Ed. Psych. 4: 159-163.) Brotvn, with 222 pupils of the sixth grade of four dif- ferent schools as subjects, found that the drill group gained 9.8 points on the second test over the first after twenty les- sons, while the non-drill group gained only 5.9 points. Ap- parently, the tests used were not standard tests, however, and no measure of reliability was applied to the data. (Jour, of Ed. Psych. 3: 485-492.) Another group of investigators, going a step further, have undertaken to evaluate special types of drill. Mead and Johnson, employing as subjects 209 pupils in grades 5 and 6, performed an experiment designed to test the relative merits of the Courtis Standard Practice mate- rial and the Thompson minimum essentials. They divided the pupils into two groups, approximately equal in size and initial achievement, as determined by the Courtis Standard Tests, Series B. Both groups were practiced ten minutes daily for ninety consecutive days, one group with the Courtis material and the other with the Thompson mini- mum essentials. Then the Courtis Tests were repeated with both groups and the amount of gain noted, with the following results : Total Gross Gain on All the Processes Group Attempts Accuracy Courtis 10.4 32 Thompson 4.8 5 Taken at their face value, these figures indicate that the Courtis material is more economical than the Thompson minimum essentials. But no measures of reliability were used to establish the validity of the data. Then, too, there may be some special transfer effect from the Courtis Prac- tice material to the Courtis Tests, Series B, which would favor the Courtis group. Again, nothing is said concerning the equality of the teachers conducting the drill in the two groups. (Jour, of Ed. Psych. 9: 287-297.) 8 A Comparative Study of Similarly, Kelly compared the progress of three groups working under different types of drill. One group, com- posed of grades 5 to 8 and numbering 133, practiced with the Courtis Standard Practice material; another group, composed of grades 4 to 8 and numbering 146, used the Studebaker Economy Practice Exercises; while a third group, composed of grades 4 to 8 and numbering 173, prac- ticed with special material prepared by the teacher. All the groups were given an initial test with the Courtis Stand- ard Tests, Series B. The practice continued for twenty days, each group devoting the same amount of time to the drill. Then the Courtis Tests, Series B, were repeated, with the following results : Gain in Per Cent Groui) Spi'i'd Accuracy Courtis 21 16 studebaker 12 11 Special material 7 10 Apparently, the Courtis material is most effective ; but no attempt was made to equalize the learning ability of the different groups or the teaching ability of the teachers con- ducting the drill in the various groups. (Jour, of Ed. Re- search, 2: 693-709.) In a similar experiment Evans and Knoche found the Studebaker Economy Practice Exercises superior to the regular class work. (Jour, of Ed. Psych. 10: 263-276.) Another noteworthy study in the field of drill is Free- mail's pedagogical experiment in handwriting to test the value of a set of principles which he had derived from an extended laboratory study of the handwriting movement. In summarizing this experiment, he wrote : "In the pedagogical experiment a system of exercises was devised to put in practice and test the foregoing principles. In these exercises emphasis was laid first upon hand posi- tion. The child was taught to hold his hand so that the wrist did not slant more than 45 from horizontal, to sup- port his hand upon the third and fourth fingers, to slide it easily upon this support, and to grasp the pen lightly with the fingers curved in a natural manner. Exercises were given to develop free and easy sideward movement of the hand across the page. The organization of the movement into units was encouraged and controlled by requiring the child to write to a count. The count was so arranged as to make the division between the units come at appropriate places in the letters. The speed of writing was gradually increased as the child acquired practice and maturity. Finally, the general educational principle that the child should have clearly in mind the aim of his practice was ap- Directed and Undirected Teaching 9 plied by directing the child's attention to the analysis of the form of his own letters and to their systematic correction." Freeman used as subjects for this experiment grades 2 to 8 in three of the Kansas City schools — namely, the Long- fellow, the Bryant, and the Quindaro. The Longfellow school served as the experimental school, and the Bryant and the Quindaro as check or control schools. The work in the experimental school was directed for a period of eight months according to the special plan de- vised by Freeman and his assistants. Throughout this same period the control schqols followed a modified Palmer System, with no systematic supervision on the part of those conducting the experiment. All the schools were rated in September on the Freeman Scale, and again in December, February, and May. The pupils of the experimental school were kept informed of their progress, and their attention directed to the analysis of their own writing habits and to the systematic correction of their errors. The control schools received no communication while the experiment was in progress, except the request for test papers at the proper times. In tabulating the results, only the records of the pupils who were present for the first and the last tests were con- sidered. The complete records were found to number 149 in the Longfellow, 194 in the Bryant, and 157 in the Quin- daro. The grading was all done by one person using the Freeman Scale. The following results were obtained: Average Gain Speed Quality Standard 7.1 1.6 Experimental School 28.0 4.34 Control Schools 20.5 0.43 The foregoing data show a decided advantage in favor of the experimental group, particularly in quality. The con- trol schools made nearly as much gain in speed, but they made less than the standard gain in quality. (Freeman, F. N., The Handwriting Movement.) A more recent study is that of O'Brien, dealing particu- larly with the development of speed in silent reading. His specific problem was : "To construct for the teacher in the classroom, types of training in effective rapid silent reading, based upon the findings of experimental science." The in- vestigation seeks to determine the factors affecting the rate of silent reading and to provide specific training for each of these factors. The chief emphasis is placed upon the fol- lowing : 10 A Comparative Study of 1. Training in rapid silent reading. 2. Training to decrease vocalization. 3. Training in perception. O'Brien used as subjects grades 3 to 8 in twenty schools, located in nine cities in Illinois. These subjects were drawn from practically every social strata in our heterogeneous population. After eliminating the records of all the pupils who were not present for all the tests, 875 complete records were obtained. Of this number, 32 were in the third grade, 263 in the fourth, 154 in the fifth, 128 in the sixth, 206 in the seventh, and 92 in the eighth. The pupils of each grade were divided into two groups approximately equal in num- ber and reading rate on the basis of the rate scores on the Courtis Silent Reading Tests. In addition to dividing each grade into two groups whose aggregate rate scores were approximately equal, the grade was subdivided into pairs of pupils approximately equal in rate of reading. Then one member of each pair was placed in the experimental group and the other in the control group. The period of training extended from April 8, 1919, to May 29, 1919, comprising thirty-nine school days. The same teacher taught both the experimental group and the control group in each grade. Both groups were given the Courtis Silent Reading Tests at the beginning, at the mid- dle, and at the end of the training period, and progress was measured in terms of the scores made on these tests. Writ- ten directions were placed in the hands of each teacher co- operating in the experiment. These directions provided for thirty minutes' daily drill in rapid silent reading, for accu- rate checking up on the daily scores, and for keeping each pupil informed of his progress. These thirty-minute pe- riods constituted the total amount of time devoted to read- ing while the experiment was in progress. The pupils were asked to read as fast as they could for two or three min- utes ; then they were asked to reproduce, orally or in writ- ing, what they had read. The reproduction was by free paraphrase or in response to specific questions based upon the text. Thus reading and reproducing alternated through- out the daily period. Each practice period had a specific aim — namely, (1) to increase the rate of reading, (2) to decrease vocalization, or (3) to increase the perception span. To insure greater accuracy, all the Courtis tests were scored by trained clerks working under the director of the experiment. O'Brien found a decided advantage in favor of the experi- mental groups. The comparative gains in rate of re^d- Directed and Undirected Teaching 11 ing, in terms of the average number of words per minute, were: Gross Percentage Experimental Groups 110.2 56 Control Groups 46.2 25 In comprehension the gains and losses were small, but in favor of the experimental groups. The results were as fol- lows: Gross Percentage Experimental Groups 4.4 0.09 Control Groups —3.6 —0.07 But no measure of the learning abihty of the two groups was taken. While they were approximately equal in initial achievement, there may have been a significant difference in learning ability. ^ The eye-movement records show that the improvement is effected by a lessening of the number of fixations rather than by a shortening of the average duration of the fix- ations. There is, also, a decrease in the number of regres- sive movements. The movements become more regular and rhythmical. (O'Brien, J. A., Silent Reading.) In the field of general method of teaching two questions particularly have been treated experimentally: 1. Which is the best method of teaching — the textbook method, the lecture method, or the laboratory method? 2. Which is the more effective method of teaching — the inductive or the deductive method? Relative to the first of these problems, Mayman made an investigation to determine the best method of teaching Physics in grades 7 and 8, using as subjects all the classes of grades 7 and 8 in Public School 84, Brooklyn. The total number of pupils was approximately 500. The experiment extended over the first and second school terms of 1911. The classes "were so divided that each lesson in science was taught to each class of the same grade by a different method. Upon the completion of the series of lessons, the classes were tested by uniform tests, and thus the effect of each method noted." ' The totals of the average percentage on these tests gave the following results : Book method 1555 Lecture method 1891 Laboratory method 1949 The laboratory method produced the highest average per- centage in sixteen out of thirty lessons, while the lecture method produced the highest average in ten out of thirty lessons. In no case did the textbook method produce the 12 A Comparative Study of highest average. The total time consumed by each method was: Lecture method, 121 minutes; laboratory method, 143 minutes ; and the book method, 232 minutes. On the basis of percental attainments, therefore, the lab- oratory method ranks first, the lecture method second, and the book method third. With respect to the time consumed, the lecture method ranks first, the experimental method second, and the book method third. But the validity of these conclusions is open to question. The experimenter's statements concerning his method and technique are vague and indefinite. Apparently, the method of checking up on the results is unreliable, being based upon teachers' marks or grades. (Jour, of Ed. Psych. 6 : 246-250.) Wiley performed a similar experiment with respect to high-school chemistry. His main problem was to deter- mine the best of three methods of teaching chemistry — namely, the textbook-recitation method, the lecture method, and the laboratory method. The subjects for this experi- ment consisted of twenty-four junior and senior high- school students. They were divided into three groups, equal in size and in ability as determined by the grades in physics. The material used was uniform, consisting in each case of a typewritten lesson in three different forms, adapted to the three different methods to be employed. One group was taught by the book method, one by the lecture method, and one by the laboratory method. Tests were given immediately after each lesson, and repeated after one week and again in four weeks. The papers were graded according to a uniform plan, each lesson having been di- vided into a number of important ideas and a group of amplifying ideas. The former were given two credits each and the latter one credit each. The lessons were so ar- ranged that each one had a credit value of 40. In scoring the papers, full credit was given for an idea correctly stated and half credit for a partially correct statement. By this method of scoring each group achieved the following re- sults : Lesson Immediate One Week Four Weeks Time Lecture 1621 1242 1102.5 27.7 min. Textbook 16.50 1452.5 1107.5 26.3 " Laboratory 1630 1290 1142.5 44 " From the foregoing data Wiley drew the following con- clusions: 1. There is no great difference in the three methods so far as imparting knowledge is concerned. 2. "For immediate learning the textbook method is un- questionably superior." Directed and Undirected Teaching 13 3. "For permanent learning the laboratory method is perhaps slightly superior." 4. "In every respect the lecture method is least effective in imparting knowledge to high-school students." 5. "Ranking the methods in economy of time, they would stand: (1) Textbook, (2) Lecture, (3) Laboratory." But the number of subjects in each group in this experi- ment was too small to yield reliable results, and the method of equalizing the groups was unsatisfactory. (Jour, of Ed. Psych. 9: 181-198.) The second problem of general method — namely, Is it better to follow inductive or deductive methods of teach- ing? — has been attacked by Sach and Winch. Sack attempted to determine the comparative merits of an inductive method of teaching logical fallacies and of the formal study of logic. Two groups participated in the ex- periment. Group I, three in number, had never studied logic ; whereas Group II, five in number, had just completed a course in formal logic. The groups were equalized in ability oh the basis of their previous school grades and the judgment of their instructors. Then Group I was given a special drill for about twenty hours in detecting falla- cious forms of reasoning. Group II had devoted about eighteen hours to this topic in their regular course. After the special drill, both groups were tested on forty-five forms of logical fallacies. In this test the subjects were asked to state with reference to each fallacy: (1) valid or not, (2) his reason for calling fallacious any argument so adjudged, and (3) the name of the fallacy. On all these points Group I excelled Group II, as shown in the following partial sum- mary of results : Validity Explanation Name Group Riciht V. G. Good Right Group I 42 24 3.6 22.6 Group II 35.3 9.8 9.3 2.3 Sach drew the conclusion that : "The learning of the va- rious forms of fallacious reasoning by an essentially induc- tive method makes a learner more efficient in the subse- quent detection of fallacious reasoning, in the assignment of reasons for such fallaciousness, and in the subsumption of the fallacies under their proper logical designations, than does the formal study of logic as commonly taught in university classes." Group I may have been favored, however, by the factor of recency ; for, apparently, Group I was drilled after Group II had completed their work. Furthermore, the method of equalizing the groups is very unreliable. Consequently 14 A Compa7-ative Study of much of the superiority of Group I may have been due to superior logical ability. Then, too, the number of subjects was too small to yield reliable results. (Jour, of Ed. Psych. 4: 381-392.) A more recent and more valuable study of this problem is Winch's Inductive versus Deductive Teaching. "The main problems were two in number. In the first place, an attempt was made to discover which of the two methods gave the better results when the children were tested on precisely what they had been taught or had learned. In the second place, an endeavor was made to find out which of the two methods gave the better results when the children were tested on new material." The pupils of standards 3 to 8 in five different schools of London acted as subjects for the experiment. They in- cluded both boys and girls, ranging in age from 8 to 15 years and numbering 193. Each standard was divided into two groups, equal in size and in initial achievement, as de- termined by preliminary tests in geometrical definition. After this division, one group was taught the definitions of geometrical forms by the inductive method and the other group by the deductive method. Some of the teaching was done by the regular teacher, some by the headmaster of a participating school, and some by the director of the experi- ment. In some instances the director taught the inductive group and in other instances the deductive group. Both groups were tested for immediate reproduction and for de- layed reproduction after one week, and again after approxi- mately one month. They were tested, also, on their ability to apply the definitions to new material. Winch found the deductive groups, on the whole, supe- rior in both immediate and delayed reproduction, and the inductive groups superior in applying the definitions learned to new material. (Winch, W. H., Inductive versus Deduc- tive Methods of Teaching.) The differences between the two groups, however, are too small to be very significant. They may be accounted for in part at least by difference in the learning ability of the two groups, for no reliable measure of learning ability was applied in dividing the pupils into two groups. Other variable factors that may have affected the results are: (1) difference in teaching ability between the various pairs of teachers when both groups were not taught by the same teacher, and (2) lack of uniformity in grading the papers. Furthermore, the same teacher may be more skillful in in- ductive than in deductive teaching, or vice versa. In like manner many other problems of method have been Directed mid Undirected Teaching 15 subjected to experimental treatment, which, for the sake of reasonable brevity, must be omitted from this review. The present discussion is far from exhaustive, but enough has been written to indicate the nature and extent of the experimental work that has been done in the field of method and to exhibit the technique commonly employed by the experimenters. With respect to method and technique employed in the foregoing studies the following criticisms apply generally, though not uniformly: 1. The experiments are too short and the subjects par- ticipating in them are too few in number to insure reliable results. 2. In most cases no measures of reliability are applied to establish the reliability of data. 3. Competing groups are not equalized in initial achieve- ment and in learning ability, or the method of equalizing the groups is unsatisfactory. No mental tests are used as a means of determining the learning ability of groups. 4. The teacher is a variable factor in most of these experi- ments, which the experimenter failed to control or to make due allowance for. Group III: Supervised Study Much has been written about supervised study, but the experimental literature in the field consists of a few arti- cles which will be reviewed briefly in the following pages. Breslich ('12) conducted the first important experimental investigation in the field of supervised study. For this experiment he used as subjects two sections of a class in algebra in the University High School of the University of Chicago. The two sections were approximately equal in achievement in this subject, as indicated by the grades for the preceding semester. These grades were distributed as follows : ABC D F Average Section A 25 25 37.5 12.5 0.0 81.4 Section B 29.4 23.5 23.5 17.7 5.9 79.4 Section B worked under the divided-period plan of super- vised study for fourteen lessons, while Section A followed the usual recitation and home-study plan. Section B did all the work in the class period of forty-five minutes ; whereas Section A devoted approximately two hours to each lesson — an hour and fifteen minutes to home study and forty-five minutes to the recitation. Upon the completion of the fourteenth lesson, the same test was given to both sections, with the following results : B c D F Average 21.4 21.4 50 62.8 6.2 37.5 25 31.9 65.5 16 A Comparative Study of A Section A 7.1 Section B 0.0 Breslich continued the experiment for six more lessons. In this part of the experiment Section A worked under the supervised-study plan and Section B under the home-study and recitation plan. On the final test, which was the same for both sections, the comparative gains over the records on the previous test were : Section A 14.7 Section B 20.9 In evaluating his results, Breslich observed : 1. Supervised study increased the percentage of lower grades and decreased the percentage of higher grades. 2. There was a slight difference in the first part of the experiment in favor of the supervised group. 3. The supervised group made more progress in learning to work independently. (School Review, 20: 505-515.) Breslich's conclusions, however, were drawn from a meager amount of data. The whole experiment comprised only twenty lessons in algebra, with two sections of one class' acting as subjects. The number of pupils participat- ing in the experiment is not stated ; but it is certain that the number was small, since only two sections of one class were employed. Furthermore, it is required statistically that the difference between two means be at least twice the probable error of the difference to be regarded as signifi- cant. More conservative practice demands that this dif- ference be four times the probable error of the difference. This test was not applied to determine the reliability of the difference in means. Since this difference was only 2.7% for the first fourteen lessons, one may well doubt whether there was any significant difference in the two groups. Moreover, in the second part of the experiment, compris- ing six of the twenty lessons, the unsupervised group, for- merly the supervised group, made a greater gain over its previous record than did the supervised group. Breslich attributed this superiority of the unsupervised group in the last part of the experiment to the effect of the supervi- sion in the first part persisting in the second part of the experiment. But the fact that this particular group, whether supervised or unsupervised, made the greater gains in both periods of the experiment leads one to suspect that its superiority was due to some factor other than su- pervision — perhaps to greater initial ability. The accep- tance of the latter explanation, which seems equally as Directed and Undirected Teaching 17 plausible as the former, would cast discredit on the whole experiment, for this difference in learning ability would be sufficient to account for the small difference between the two groups. Mimiich ('13) conducted an experiment designed to test the value of supervised study in plane geometry. The sub- jects were thirty-six pupils in the high school of Blooming- ton, Ind. They were divided in two groups of equal size and of approximately equal ability on the basis of the grades made in algebra for the three previous semesters. These students were just beginning their work in plane geometry. One group worked under a supervised-study plan and the other under the ordinary recitation plan. The unsuper- vised group prepared their lessons at home and recited dur- ing the first class period ; the supervised group recited the second class period, remaining the third period to prepare the work for the next day. If any student completed the assignment before the end of the study period, additional work was provided for him. An effort was made to teach the pupils how to study. They were free to call for help in the study period, but help was given in the form of ques- tions and suggestions. Records were kept of the grades on the daily recitations and of the number of recitations made weekly by each group, and curves were plotted based upon these records. Tests were given at the close of the first and second six weeks, covering the work done in these periods. At the end of the semester a final test was given on the work of the whole semester. Occasionally during the progress of the experiment tests were given on new material. The same test, in every instance, was given to both groups. With the exception of the final test, the quan- tity of work did not affect the grade, for the papers were graded only on the quality of the work done. All the pa- pers were graded by the instructor, but they were arranged so that he did not know whose paper he was grading. Minnich found that, on the basis of the weekly averages, the supervised group had the higher averages for ten weeks, the same for three weeks, and lower for two weeks. In the fifteen weeks the supervised group made 466 recitations,, while the unsupervised group made only 352. The two groups made the following records on the tests : Supervised Unsupervised Difference First six weeks 77.3% 68.7% 8.6% Second six weeks 81.2% 80.4% .8% Final test 92.4% 80.1% 12.3% On the tests on new material the supervised class aver- 18 A Comparative Study of aged 13.1% more than the unsupervised class. All the pu- pils in the supervised group passed, but two of the unsu- pervised group failed. The foregoing experiment extended over a period of fifteen weeks, which is probably a sufficient length of time to produce important results. The two classes were com- parable in size, and perhaps also in ability, although semes- ter grades are an unreliable measure of ability. The com- parisons were made on the basis of the weekly averages of grades made in the daily recitations, the number of recita- tions made per week, and the average scores on the special tests devised by the teacher. But teachers' marks have been found to be extremely variable, and consequently unreliable measures of progress. The investigations of Kelly, Starch and Elliott, and Johnson furnish convincing evidence of the unreliability of the marks assigned by teachers. (See Kelly, F. J., Teachers' Marks; School Review, 20: 442-457; 19: 13-24; 21: 254-259; 21: 676-681.) Moreover, compar- ative scores are often influenced largely by the teacher's attitude. A teacher who believes in supervised study will do better teaching with the supervised group, and may un- consciously favor this group in the grading of papers. In view of these facts, Minnich's results cannot be accepted at their face value. Moreover, his statistical treatment of his data is inadequate, for he failed to apply any measures of reliability. Furthermore, his report does not provide sufficient data to enable one to apply the measures of relia- bilitv and thus estimate his work. (School Review, 21: 670-675.) Grace A. Dunn ('17) conducted an experiment in super- vised study in language, using as subjects the pupils of the fourth grade in a town system near New York City. The pupils, numbering twenty-two, were divided into two sec- tions, known as A and B, of equal size and of approximately equal ability, as determined by the scores on the Trabue and Thorndike Tests in Language. Both sections were given twenty lessons of thirty minutes each. They had the same assignment and the same amount of time for study; but the study of Section A was closely supervised, while Section B was left to its own resources in preparing each lesson. At the close of each week the pupils were tested and rated on their work for the week. The tests were the same for both sections. The general average of these rat- ings for the four weeks gave the following results : Section A, the supervised group_ 85% Section B, the unsupervised group 43% Directed and Undirected Teaching 19 On the basis of these averages, Miss Dunn concluded that there was a decided difference in favor of supervised study ; but she failed to use any measures of reliability, and her method of grading was not clear. It appears, however, from her description of her grading that her ratings were based upon tests and exercises making a heavy demand upon judgment. For example, in speaking of the work of the first week, she states: "In rating the work, the teacher considered the organization, originality, clarity of expres- sion, and general appearance of the reproduction work, and the intelligent appreciation and interest which should have been the outcome of the study of the poems." Now, it is obvious that the foregoing factors could not be rated with a reasonable degree of accuracy. Consequently allowance must be made for large errors in grading. Moreover, there is nothing to indicate that the teacher made any attempt to maintain an unbiased attitude in conducting the experi- ment. On the contrary, from reading the report, one gets the impression that she slighted the work of the unsuper- vised section and put forth every effort to achieve unusual results with the supervised section. The very low general average of the unsupervised section confirms this impres- sion. Again, all measures of reliability are omitted. Only average scores are included in the report. Obviously, ex- periments conducted in this manner have little, if any, sci- entific value. (Teachers' College Record, Vol. 18: 430- 437.) Breed ('19) conducted an experiment to determine the value of the divided or double-period plan of supervised study. He used as subjects thirty-four classes in Latin, Algebra, and English composition from fourteen different high schools. Satisfactory records were obtained from 596 pupils. Definite and detailed directions in mimeographed form were placed in the hands of the principals and teachers co- operating in the experiment. Two plans were outlined, either of which a school might follow. Plan I provided for one teacher to teach two sections approximately equal in size and ability as determined by preliminary tests devised by the teacher or by the mid- year marks in the subject chosen for the experiment. These sections were known as Section A and Section B. Section A was taught by a supervised-study plan for six weeks, and Section B by the usual recitation plan. Both sections had the same length of period and covered the same subject- matter both in kind and amount. Thus the work continued for six weeks ; then both sections were given the same semi- 20 A Comparative Study of final test devised by the teacher. The procedure was then reversed by teaching Section B according to the supervised- study plan, and Section A by the usual recitation plan. Again, both sections covered the same subject-matter and took the same final test devised by the teacher. All the test papers were scored first by the teacher; then they were sent to the Educational Laboratory of the University of Michigan and scored by a competent grader selected by the director of the experiment on the basis of special fit- ness and paid for his services. The same grader in every case scored all the papers of the two comparable classes. The average of the teacher's grades and the laboratory grades was taken as the true score. Plan II required two teachers, known as X and Y, and two sections, known as Section A and Section B. Teacher X taught Section A by the supervised-study plan for six weeks, while Teacher Y taught Section B by the usual reci- tation plan. After the semifinal test, which was devised by the teachers in cooperation and was the same for both sections, the teachers exchanged sections for the final pe- riod of the experiment. At the close of the second period the same final test, devised by the teachers in cooperation, was administered to both sections. All the scoring was done in the same manner as in Plan I. All but one school chose Plan I, probably because this plan could be more readily adjusted to existing school conditions. A list of suggestions designed to embody the theory and practice of the divided-time scheme of supervised study was placed in the hands of each cooperating teacher. These suggestions emphasized particularly the following factors : Individualizing instruction ; Directing the pupil's efforts ; Helping the pupil judiciously; Teaching the pupil how to study ; Increasing the importance of the assignment ; Special attention to the lowest third of the class ; Providing extra work for the highest third ; Decreasing the amount of home study. In general, the teachers were directed to proceed in the following manner : "The pupils have begun to work. The getting-ready process must be prompt and precise, with a minimum of noise. No questions are permitted audibly, but by raising the hand the pupil indicates a desire to consult the teachers, who pass quietly to the pupil's desk. The pupil and teacher in whispers confer on the diflficulty. When not so occu- pied, the teacher moves quietly up and down the aisles, stop- Directed and Undirected Teaching 21 ping at each desk to inspect each pupil's work. When she finds a pupil employing a wrong method, she stops and in a low tone asks why he uses this method. She always re- quries the pupil to give a reason for what he does. The pupil is expected to make his own corrections with mini- mum of suggestion from the teacher. When absolutely nec- essary to do so, she will give the required information." Furthermore, each teacher was provided a mimeographed copy of Whipple's Rules for Study to be used in teaching the pupils how to study. Comparisons were made on the basis of both coyitinuous and more strictly comparable groups. The continuous groups were composed of all the pupils present for all the tests. The strictly comparable groups were selected from the continuous groups in such a manner that the groups to be compared were of the same size, matched individually in ability, and equal, therefore, in average ability. Moreover, these comparable groups were taught by the same teachers and belonged to classes of approximately equal enrollment. Averages were computed: (1) on the basis of the teachers' scores, (2) on the basis of the laboratory scores, and (3) on the basis of the combined scores. But the final judg- ments were based on the combined scores, and the progress of the various classes was measured by the increases in the average scores. The results obtained by Breed are sum- marized in Table I. TABLE I Showing the Average Percentage of Gains and Losses for the Continuous and Comparable Groups Peircentage of Gain or Loss No. of Pairs Preliminary Semifinnl Subject of Classes to Semifinal to Final Continuous Groups Algebra 7 —0.2 —5.3 Latin 2 9.0 7.1 English Composition 2 — 4.4 — 3.9 Comparable Groups Algebra 7 1.0 2.7 Latin 2 3.7 5.4 English Composition 2 — 3.8 — 6.1 In this table the percentages are all in terms of super- vised study, and each represents loss or gain in a particu- lar subject and for one period of the experiment. If the percentage is plus, it means that supervised study proved to be superior; if it is minus, it means that supervised study proved to be inferior to the usual recitation plan. 22 A Comparative Study of In the three subjects the poorer pupils gained on the average 3.4%, and the better pupils lost 5.7%. Breed was unwilling to hazard an interpretation of his data, but he expressed his opinion to the effect that : "Supervised study of the kind tested facilitated the prog- ress of the poorer pupils, and to a correspondingly greater degree retarded the progress of the better pupils." "Until the divided- and double-period plans meet the ob- jections urged against them, a selective or differential plan of study supervision should be favored by secondary schools — that is, a plan which concentrates on the pupils of less ability." "The technique of supervising the poorer pupils needs to be improved ; a technique of supervising the study of the brighter pupils needs to be developed." The foregoing experiment is by far the most extensive and valuable of all that have appeared in this field. The method of conducting the experiment is as scientific as it could well be made under real school conditions. An at- tempt is made to control the variable factor of the teacher by providing for an exchange of teachers at the middle of the period or for one teacher to teach both sections. The essential conclusions, however, are based largely on central tendencies, which, taken alone, are inadequate measures of group progress. All measures of reliability are omitted, and not enough data are provided to enable one to apply these measures and thus determine the validity of the re- sults. The advisability of exchanging groups at the mid- dle of the period may well be doubted, for this allowed only six weeks for any one group to work under supervised study. It is probable that six weeks is too short a time to produce any marked differences in such subjects as algebra, Latin, and English composition. Further, the experiment deals with only one type of supervised study, and the in- structions to the teachers indicate that , the supervision tested consists largely of better assignments and more in- dividual help, particularly for the poorer pupils. Then, too, there is no statement to the effect that the teachers cooperating in the experiment had any special training in the theory and technique of supervised study previous to their participation in this experiment. On the other hand, no doubt, they had considerable training for the ordinary recitation plan. It is apparent that this lack of training, if present, would place the supervised study group at a con- siderable disadvantage. This experiment, therefore, should not be accepted as final, even with respect to the type of supervision tested ; rather it indicates, as Breed points out, Directed and Undirected Teaching 23 that the technique of supervising study needs to be devel- oped and improved. (School Review, 27: 186-204 and 262- 284.) From the foregoing studies it is obvious that direction of the pupil's study has not up to this time yielded any very valuable results, but the data in the field are meager and unreliable. There is urgent need for more experimental investigations of supervised study, and for a refinement of the technique of directing the learning activities of pupils. A very recent study, which cannot properly be classified in any of the preceding groups, is Pit?nan's The Value of School Supervision. The problem of this study is very sim- ilar to the problem of the present investigation. Pitman undertakes to measure the value of school supervision on the basis of quantitative data. His specific problem is : "What is the effect of supervision upon the work of rural schools when the supervision is done according to the Zone Plan?" The Zone Plan is defined as: "A plan of supervi- sion in which the supervisor divides his entire supervisory district into territorial limits, each of which serves as the territorial limits for one week of supervisory effort." Two groups of schools located in Brown County, S. D., were selected for this experiment. The supervisory zone, constituting the experimental group, was composed of fif- teen schoolrooms, and the control group of twenty-five schoolrooms. The two groups of schools were approxi- mately equal in the length of the school terms and the social and economic status of the communities. With respect to years of training beyond the eighth grade and years of ex- perience in teaching, the two groups of teachers compared as follows : Training Experience M'-dian Median Experimental Group 4.2 2 Control Group 4.36 2.33 In the matter of certification the experimental group had a slight advantage, as shown by the following figures : First-Grade Second-Grade ' Percentage Pcrcey^tatje Experimental Group 46 54 Control Group 36 64 All the pupils in grades 3 to 8 in the experimental group of schools, numbering 114, were used as subjects. On the basis of the scores made on standard tests in reading, spell- ing, composition, handwriting, and arithmetic, a group of 114 pupils were selected from the larger control group, num- 24 A Comparative Study of bering 225, which was approximately equal to the experi- mental group. Both groups were given the same initial standard tests be- tween September 28 and October 10, 1919. The tests were ad- ministered by the director of the experiment and three com- petent assistants, the same persons testing the same schools in both the initial and final tests. The scoring was all done by members of the classes in tests and measurements in the normal school at Aberdeen, working under the direction of their teacher, Miss Ivy Husband. A different group of students scored the final tests from that which scored the initial tests, but the second group worked under the direc- tion of Miss Husband, as did the first group. The final tests were given between May 3 and May 14, 1920, and in the same order as the initial tests. The supervisor spent one week each month in making a tour of his zone. On these tours special emphasis was placed on two general types of work — namely, the super- vision of instruction and the stimulation of the social and economic life of the community. The activities which had for their immediate purpose the improvement of the character of the school work comprised : 1. Making a calendar for the events of the year, show- ing the dates of the various tours, with the special work of each tour, and the time and place of the group meetings for the teachers of the zone. 2. Writing a supervisory letter to each teacher in the zone, about one week in advance of each tour, announcing the llr.ie of the supervisor's visit and indicating the sub- jects that would be emphasized on the tour. 3. Visiting the teachers once each month in their class- rooms and observing the work of a few classes. On these visits the supervisor took brief notes on the lessons observed and made short talks to the pupils. The conferences with the teachers on these visits were usually brief, dealing only with matters of local and immediate importance. 4. Holding group teachers' meetings, consisting of the following activities : (1) A teaching demonstration by the supervisor. (2) A conference based upon this demonstration. (3) Teaching demonstrations by two classroom teachers. (4) A conference based upon these demonstrations. (5) A summary discussion of the subject under investi- gation. (6) The giving to the teachers of a list of suggestions for their guidance for the next month. These suggestions dealt with the teaching of one subject, and contained a list of Directed and Undirected Teaching 25 books from which the teachers were to select books to be read during the month. In order to stimulate the community life, and thus win for the school a more intelligent and whole-hearted support, activities were carried on along the following lines : 1. Visits in the homes of the people, the aim being to eat a meal or sleep at the home of every child in the zone. About 50% of the homes in the zone were reached in these visits. 2. The school newspaper for the zone, containing news items, the aims for the year's work, the result of the initial survey, and the account of the teachers' meetings. 3. Social phases of the teachers' meetings. To these meetings the people of the community came and brought their dinners. The afternoon was devoted to the pleasure and inspiration of the community. 4. Health and happiness meetings, in which important health subjects were discussed by competent speakers. 5. Spelling matches for socializing purposes and to pro- vide practice in spelling certain agricultural terms and words which every one should know how to spell. 6. A crusade against gophers, organized and conducted by the school. The progress of the two groups of pupils was measured in terms of improvement as shown by the scores on the initial and final tests. Pitman found that the median prog- ress of the experimental group, when expressed in terms of the progress of the control group, was 193.75%. By the equated difference method he found that it would take the control group .942 of a year to reach the point of improve- ment attained by the experimental group. The results, therefore, show a decided advantage in favor of the supervised group. But one does not know which of the agencies employed was of most value in producing the results. There is need for experimentation to determine the value of each of the agencies that are commonly em- ployed in supervision. The present investigation is de- voted to the study of one of these agencies — namely, di- rected teaching. II. Statement of the Problem Does the supervision of instruction increase to any ap- preciable degree the efficiency of teaching? Can the ef- fectiveness of supervision be measured quantitatively? This study undertakes to answer these questions for one particular phase of supervision, which may be designated as directed teaching. The specific problem may be stated 26 A Comparative Study of thus: How effective are definite directions as a device for the supervision of instruction? Directed teaching, as used throughout this study, refers to that particular type of supervision in which the teacher and the supervisor, in cooperation, devise a specific plan for the direction of the work of a class in a particular sub- ject. Then the teacher instructs the class according to this plan. In the meantime the supervisor keeps closely enough in touch with the class to be able to offer helpful suggestions from time to time and to make constructive criticisms of the teacher's work. CHAPTER II THE GENERAL PLAN OF THE STUDY This investigation attempts to determine the value of directed teaching by measuring the progress of comparable groups of pupils, one group working under a directed-teach- ing plan and the other without any such plan. The experi- ments have been conducted, however, under the usual school conditions. Consequently many compromises have been necessary between the precision of science and the limita- tions of school conditions, for, under school conditions, it is not possible, as a rule, to control rigidly all the factors that may affect the results. Nevertheless, an effort has been made to control these factors as far as possible, and to make note of all the variable ones so that due allowance may be made for them in interpreting the results. Two gen- eral plans have been followed in conducting the experi- ments, which may be outlined briefly as follows : Plan I 1. Two teachers, equal, as nearly as possible, in teaching ability. 2. Two grades, known as A and B, and approximately equal in size, ability, and initial achievement in the subject. 3. The initial achievement of the two grades, to be deter- mined by standard tests administered by the principal or by the writer, all the scoring being done by the writer. 4. One teacher directs the work of Grade A according to a directed-teaching plan, devised in cooperation with the writer. The other teacher instructs Grade B, relying upon her own resources and the help given by the regular super- visory force. Aside from the specific directions placed in the hands of the one teacher, both are subject to approxi- mately the same amount and kind of supervision. 5. Both grades cover the same subject-matter, both in kind and amount. 6. The progress of the two grades is compared on the basis of the scores made on the standard tests given again at the close of the experiment. These tests are adminis- tered and scored in the same way as the initial tests. Plan II 1. One teacher. 2. Two sections of the same grade or high-school class, known as Section A and Section B, approximately equal in 28 A Comparative Study of size, mental ability, and initial achievement, as determined by standard tests or special tests devised by the teacher. 3. The teacher directs the work of Section A according to a directed-teaching plan, and teaches Section B, follow- ing her accustomed method. 4. Both sections cover the same subject-matter, both in kind and amount. 5. Both sections are given the same tests again at the close of the experiment, and the progress of the two groups compared on the basis of these scores. 6. In case tests devised by the teachers are used, three tests are given on different days, and the average made on the three tests taken as the true score. The questions used in these tests call for definite answers, so that they can be graded with a fair degree of accuracy. They are graded first by the teacher, then by the writer, and the average of the two ratings taken as the true score. A Specific Plan for Each Experiment Each teacher working under direction devised, in cooper- ation with the writer, a specific plan for the direction of her group. An attempt was made to embody in these plans the best available knowledge of how to teach and study the subject under consideration. To this end constant refer- ence was made to the best books, and to original investiga- tions dealing with the learning of the subject. For exam- ple, the plan for directing the teaching of handwriting was based largely on the following : Freeman, F. N., The Teaching of Handwriting. Freeman, F. N., The Handwriting Movement. Monroe, W. S., Measuring the Results of Teaching. Nutt, H. W., Rhythm in Handwriting (Ele. Sch. Jour., 17: 432-45), Graves, S, M., A Study in Handwriting (Journal of Edu- cational Psychology, 7: 483-494). Freeman, F. N., Practical Studies of Handwriting (Ele- mentary School Teacher, 14: 167-179). The plans for other subjects were based upon similar sources. In order that the reader may know more definitely the nature of the specific directions given to the teachers, a number of these plans will be reported in full in the discus- sions of the experiments in the next chapter. It is freely admitted that these plans for directing the teaching are imperfect. They represent in each case what the cooperating teacher and the writer judged to be the best practice in the eff'ective direction of a class. In so far as they fail to embody the best practice, the effect would be to Directed and Undirected Teaching 29 reduce somewhat the final scores of the pupils working un- der the directed teachers. Variable Factors 1. The Teacher. — In both the foregoing general plans the teacher is a variable factor. It is highly improbable that any pair of cooperating teachers in any one experiment would be exactly equal in teaching ability, or that they would teach with an equal degree of interest and enthu- siasm throughout the experiment. Furthermore, it seems to the writer just as improbable that any teacher would teach two sections or classes with an equal degree of effi- ciency and at the same time follow an essentially different method with each group. In fact, the teachers working under Plan II (see page 27) confessed their inability to keep the two methods distinct, although they made an ef- fort to follow instructions implicitly. In selecting teachers to cooperate in the experiment, how- ever, care was taken to avoid favoring constantly either the directed or undirected group. Teachers had to be selected, in part, on the basis of their willingness to participate in the experiments ; but, as far as possible, teachers holding similar positions in the same school were paired one against the other. This arrangement provided for approximately equal help from the regular supervisory force. It was found to be impracticable, under school conditions, to equal- ize the factors of education and experience, because teach- ers holding similar positions in the same school system vary widely in their preparation and experience. These factors, however, should have due consideration in estimating the comparative efficiency of teachers. But the reliability of education and experience as meas- ures of teaching efficiency is still an open question. After studying 504 cases, Meriam came to the conclusion that: "After the first year the amount of experience is not an im- portant criterion for efficient teaching in the elementary schools." In this study the teachers were given an effi- ciency rating by their principals, which was based largely upon subjective factors. On the basis of an extended study of the relation of length of service to efficiency in teaching, Coffman came to a similar conclusion — namely: "These ta- bles and graphs accompanying them certainly justify the conclusion that experience does not contribute much to the efficiency of the public school teacher." Similarly, Coff- man found no uniform tendency existing between the amount of salary received and education beyond the ele- mentary school. His essential conclusions on this point 30 A Comparative Study of were: (1) "The first four years of training beyond the elementary schools have little or no effect on salary; (2) correlation between salary and education becomes increas- ingly marked with' each succeeding year after the fourth year." Using the salaries received by men in the private secondary schools of five cities as a measure of efficiency in teaching, Thorndike found that the full effect of the expe- rience in teaching in private secondary schools is reached in three years. For the full text of these investigations, see: Meriam, J. L., Normal School Education and Efficiency in Teaching. Coffman, L. D., The Social Composition of the Teaching Population. Thorndike, E. L., The Teaching Staff of Secondary Schools in the United States (Bureau of Education Bulle- tin, 1909, No. 4). These studies, then, indicate that experience after the first three years does not contribute much to eflficiency in teaching. They suggest, also, that education, particularly beyond the elementary school, is a more reliable measure of teaching efficiency than is years of experience. These facts should be kept in mind in considering the comparative efficiency of the various pairs of teachers cooperating in the various experiments of this study. The data bearing on the efficiency of teachers will be presented in connection with the report of each experiment. 2. The Pupils. — The pupils participating in these experi- ments constitute a second variable factor. An effort was made to select groups of pupils that would be comparable in size, in mental ability, and in initial achievement in the subject ; but it was not possible, under the limitations im- posed by school conditions, to arrange strictly comparable groups. An attempt was made, however, to determine the native ability and initial achievement of each group by em- ploying the devices described in the following paragraphs : (1) Native Ability. — In most cases the group intelli- gence tests were used to determine native ability. These tests were employed particularly in the groups that were to study the subjects that have been found to have a marked correlation with general intelligence. In subjects, such as handwriting, which have been found to have only a very low correlation with general intelligence, secondary meas- ures have been employed — namely, average age and social composition of the pupils. (2) Initial Achievement. — Initial achievement was de- termined in each case by preliminary tests, which were the Directed and Undirected Teaching 31 same for groups to be compared. The standard tests were used in all cases, if such tests were available. If such tests were not available, special tests devised by the teacher were substituted for the standard tests. The questions used in these special tests were so framed that definite answers were required, and a set of directions for scoring the pa- pers were written out, thus securing reasonable uniformity in grading the test papers. A certain number of points was allotted to each correct answer and partial credit for incomplete answers as designated in the written directions for scoring the papers. Each pupil was given a definite rating on this point scale. Three fifteen-minute tests were given to each group on three successive days, and the aver- age score on the three tests taken as the true score for each pupil. Finally, in order to minimize still further the influence of the teacher in determining class scores through the man- ner of giving the tests and scoring the papers, the initial and final tests were given by the writer or the principal of the school. Moreover, all the papers were graded by one other competent grader, and the average score taken as the true score. In general, an attempt was made to plan the experiments so that neither of the groups to be compared would be fa- vored constantly by the factors that could not be rigidly controlled. Each experiment will be reported separately in the following chapter, and any significant variations from the general plans outlined in the present chapter will be noted in these separate reports. In each of the frequency curves of the following chapter the horizontal line indicates, by intervals, the scores made on the test, and the vertical line indicates the number of pupils making each score. In plotting the frequency curves, it is assumed that all the scores in any interval are concen- trated at the mid-point of that interval. Consequently the height of the curve at any mid-point shows the number of pupils falling in that interval. In each graph the solid line represents the group taught by the directed teachers, and the broken line the group taught by the undirected teachers. In each table, or fre- quency distribution, the A-groups are the directed groups and the B-groups the undirected. The terms directed and undirected are applied conven- iently, to the respective groups of pupils as well as to the teachers. A directed group, for example, means a group taught by a directed teacher ; an undirected group, one taught by an undirected teacher. CHAPTER III THE EXPERIMENTS The present chapter is devoted to a presentation of the experiments, all of which were performed in the school years of 1919-1920 and 1920-1921. Each experiment is ac- corded a separate and detailed treatment, then all the re- sults are brought together in summary form in the final chapter. All the following experiments have the same aim — namely, to determine whether directed teaching is more effective thayi undirected teaching. Since this is the aim of all the experiments, it need not be repeated in the report of each experiment. Experiment 1 The Fundamental Operations of Arithmetic The Subjects. — The subjects for this experiment con- sisted of forty-five pupils from the fifth grade of a small town school. They were divided into two sections, known as Section A and Section B, numbering twenty-three and twenty-two, respectively. With respect to age and nation- ality, the two sections were nearly equal, as shown by the following comparison : Per Cent of Av. Age A.D. Native Parentage Section A 11.28 . 1.09 100 Section B 11.15 .73 95 On Monroe's Standardized Reasoning Test in Arithmetic the two sections made the following scores : Correct Principle Correct Answer Section A 18.5, P. E. 4.5 10.5, P. E. 2.8 Section B 18.7, P. E. 4.9 11.5, P. E. 3.0 Difference .2 1.0 P. E. of Difference 1.4 .87 The two sections, therefore, were approximately equal in the ability to solve problems, with a slight advantage in favor of Section B. The Teacher. — Two teachers, known as X and Y, cooper- ated in this experiment. These teachers were holding sim- ilar positions in the school system and were both receiving the same salary. Teacher X was a normal-school graduate, with one year of experience in teaching; while Teacher Y had only an elementary-school education, with two summer terms of special training for intermediate-grade teaching, and had eight years of experience in teaching. In so far as Directed and Undirected Teaching 33 teaching ability can be measured in terms of salary, edu- cation, and years of experience, it would seem that Teacher X had a slight advantage over Teacher Y, having had about five years in school beyond the elementary school. On the other hand. Teacher Y had a considerable advantage from the standpoint of years of experience. Method of Procedure. — This experiment was conducted according to Plan I. (See page 27.) The whole group was given the Courtis Standard Tests, Series B, and Monroe's Standardized Reasoning Tests in Arithmetic. These tests were administered and scored by the principal of the school. On the basis of the scores made on these tests the group was divided into two sections approximately equal in size and initial achievement. After this division. Teacher X taught Section A, following a directed-teaching plan, and Teacher Y taught Section B, following her accustomed plan. The following written directions were placed in the hands of the teacher directing Section A, the directed-teaching group : 1. Provide a twelve-minute daily practice period. 2. Use as practice material the Studebaker Practice Ex- ercises. 3. Read carefully the Teachers' Manual for the Stude- baker Exercises and follow the instructions found therein. 4. At the beginning of the experiment inform each pupil of his standing in comparison with the standards and with the other members of the class. 5. Lead each pupil to set up definite standards to be at- tained by the end of the experiment. Teacher Y taught Section B in her usual way, using for practice the textbook material and special material pre- pared from day to day. She was fully aware of the nature of the experiment, and put forth every effort to achieve good results with her class. This section, also, had twelve minutes daily for practice. The experiment continued for a period of six weeks. At the end of this time the principal of the school gave to both sections the Courtis Standard Tests, using the same form as in the initial test, and scored and tabulated the results. Results. — In tabulating the results, the records of all the pupils were eliminated who were not present for both the initial and final tests. Only two records, however, had to be eliminated from Section A and two from Section B. This left twenty-one records in Section A to be compared with twenty in Section B. The two sections were still ap- proximately equal in attempts, but Section B was slightly 34 A Comparative Study of superior in rights. In subtraction and multiplication Sec- tion B showed considerable superiority in rights. How the two groups compare on the initial and final tests is shown in the following tables and graphs : TABLE II Showing the Relative Acj hievement of Initial Test attempts THE Two Sectio NS ON THE Section A Section B Difference _-. P. E. of D.__ Addition Med. P. E. .__ 5.5 1.3 -_ 6.3 1.7 .-_— .8 --_ .47 Subtraction Med. P. E. 7.9 2.4 6.8 .8 1.1 .53 Multiplication Med. P. E. 5.4 1.1 4.5 1.1 .9 .34 rivisioN Med. P. E. 3.9 1.2 3.8 1.0 .1 .34 Section A Section B Difference __. P. E. of D.__ --_ 2.6 1.2 .__ 2.9 1.8 -__— .3 _-_ .48 rights 2.9 2.5 5.0 1.6 —2.1 1.28 2.3 3.5 —1.2 .93 1.3 1.4 1.5 1.6 1.6 .8 — .1 .93 The probable errors (P. E.'s) are all computed by the formula: P. E. equals 0.6745 times the square root of the sum of the deviations squared, divided by the number of measures, minus one, or .6745 sigma. Directed and Undirected Teaching 35 Addition At. Rt. Subtraction At. Rt. Multiplication At. Rt. Division At. Rt. 8 7 \ u' ■ ■ " / / V. M 6 5 4 c ? \ \ ^ \ \ \ - -\ - \ \ \ \ \ / / ' I J ' 111 / /' 1 ' \ \ \ — L .\. : \ \ --W A----\ p -- \\ il v-"/^ 3 \ \ ij f"Y' -- / ) — V" \\ 2 \\ %■ 1 Figure 1 — Showing the relative attainment of the two sections on the initial test. The curves are based on the median scores. o -(^ Section A o Q_ _ _ _ _ Q_ _ _ - -Q Section B Standard Scores 36 A Comparative Study of TABLE III Showing the Relative Median Attainment of the Two Sections ON THE Final Test Addition Med. P. E. 1.8 1.3 ATTEMPTS Subtraction Med. P. E. Multiplication Med. P. E. Division Med. P. E. Section A 8.5 Section B 6.0 Difference 2.5 P. E, of D .48 7.6 7.5 .1 .41 RIGHTS 1.4 1.3 Section A 5.4 Section B 3.7 Difference 1.7 P. E. of D .51 1.8 1.5 5.6 5.3 1.9 1.6 .54 6.4 5.7 .7 .36 3.4 4.7 -1.3 .43 1.3 1.0 1.3 1.5 7.3 5.0 2.3 .54 4.8 2.6 2.2 .57 1.9 1.6 2.0 1.7 Addition At. Rt. Subtraction At. Rt. Multiplication At. Rt. Division At. Rt. y 8 K ) 7 N \\ Y \ b \V 1 1 ^ o __( v r xf^,"? x\\ y-"^ 1 1 I I-S. \ ^ ., _i.4 4 ... / 1 If ---l-f- 1 1 \--^^ 3 ... ) V in ■r/j --\\ •» \ \ ^ \ J "* Figure 2. — Showing the relative attainment of the two sections on the final test. The curves are based on the median scores. (See Table III.) "O O Section A o O _ ~ Q- Q Section B — . — — ' Standard Scores Directed and Undirected Teaching 37 TABLE IV Distribution of Total Attempts and Total Rights on Both the Initial and Final Tests initial test final test attempts rights attempts rights Score 5A 5B 5A 5B 5A 5B 5A 5B 39-41 36-38 3 33-35 1 4 1 30-32 12 1 4 2 2 27-29 1 2 3 3 24-26 2 2 5 2 1 21-23 5711 3412 18-20 6422 23 04 15-17 3323 1333 12-14 12 2 7 13 9-11 12 1 4 4 6-8 2 3 2 2 3-5 4 2 2 1 0-2 3 1 Total 21 20 21 20 21 20 21 20 Median 20.3 21 8.5 13.3 30.4 24 16.5 15 P. E. 4.3 4.1 5.4 3.7 4.5 2.9 6.8 3.8 Difference —.3 —4.8 6.4 1.5 P. E. of D 1.31 1.44 1.17 1.71 9-11 12-14 15-17 18-20 21-23 24-26 27-29 30-32 33-35 36-38 Figure 3 — Showing the relative achievement of the two sections on the initial test on attempts. Curves based upon the total scores on the four fundamental operations. 9-11 12-14 15-17 18-20 21-23 24-26 27-29 30-32 33-35 36-38 Figure 4 — Showing the relative standing of the two sections attempts, based upon total scores, as in Figure 3. on the final test on 38 A Comparative Study of 27 30 33 Figure 5 — Showing the relative achievement of the two sections on the initial test. Curves based upon the total scores for rights on the four fundamental operations. Figure 6 — Showing the relative standing of the two sections on the final test, based upon total scores for rights, as in Figure 3. TABLE V Showing the Gross Gain and Percentage of Gain of the Two Sections GROSS gain percentage OF GAIN attempts rights attempts RIGHTS ABA B A B A B Addition 3.0 —.3 2.8 .8 54 —5 107 28 Subtraction —.3 .7 2.7 .3 —4 10 93 6 Multiplication 1.0 1.2 1.1 1.2 19 26 48 34 Division 3.4 1.2 3.3 1.0 87 31 220 62 Total 7.1 2.8 9.9 3.3 Av. 39 16 117 33 TABLE VI Showing the Relative Amount of Gain of the Two Sections in Terms of School Years years of gain attempts rights A B A B Addition 2.0 —0.2 2.8 0.8 Subtraction __L —0.2 0.4 1.8 0.2 Multiplication !___ 0.7 0.8 0.7 0.8 Division 2.3 0.8 1.7 0.5 Average 1.2 0.45 1.75 0.58 Final Test Gain 71.4 28.6 57.1 23.8 Directed and Uridirected Teaching 39 TABLE VII Showing the Percentage of Section A Equaling or Exceeding THE Median of Section B Initial Test Attempts 1 42.8 Rights 33.3 Experiment 2 Handwriting The Subjects. — All the pupils of the sixth, seventh, and eighth grades of four different schools, numbering 346, par- ticipated in this experiment. The pupils of two of these schools constituted the directed groups, while the pupils of the two remaining schools served as control groups. The directed groups attended schools located in towns of about 2,500 inhabitants, largely of native American stock. The undirected groups attended a school located in a town of about 1,500 inhabitants and a ward school located in the resident section of a small city. How these pupils com- pared with respect to average age and nationality may be seen from an examination of the following table : TABLE VIII Showing the Enrollment, Average Age, and Nationality of the Pupils Participating in This Experiment Directed Groups Undirected Groups Per Cent Per Cent Enroll- Average of Foreign Enroll- Average of Foreign Grade ment Age A. D. Parentage ment Age A. D. Parentage 6 73 12.7 1.0 3 51 11.7 .95 2 7 54 13.1 .95 6 53 12.6 .98 2 8 60 13.9 .99 2 55 13.3 .88 2 The Teachers. — The teachers cooperating in this experi- ment were all the regular teachers, none of whom had taken any special training for the teaching of handwriting. How nearly the two groups were equal in teaching ability could not be determined with any reasonable degree of accuracy. Only rough measures of teaching efficiency were available, such as education and experience in teaching. How the two groups compared with respect to these measures may be seen by an inspection of Table IX. 40 A Comparative Study of TABLE IX Showing the Relative Amounts of Education and Experience of THE Two Groups of Teachers Directed Undirected Years of Years of Grade Education Experience Education Experience f, < Elementary 14 Normal 3 ^ } High School 15 Normal 3 rj \ Normal 1 Normal 10 ( Elementary 10 Normal 7 Q f Normal 3 Normal 10 \ Elementary 34 Normal 11 The undirected teachers, as a group, had more education and less experience than the directed teachers. Since expe- rience beyond three years has been found to add very little to teaching ability, it seems probable that the control group was superior to the experimental group of teachers. (See page 29.) Method of Procedure. — The four cooperating schools were first studied with respect to the age and nationality of the pupils ; then, on the basis of this study, two of the schools were selected for the directed groups and the other two for the undirected groups, known as the A-groups and the B- groups, respectively. In making this division, the grades having the highest average age and the highest percentage of f -^reign parentage were placed in the A-groups, thus giv- ing a slight advantage to the B-groups. (See Table VIII.) When this division had been made, it was found that the teachers of the B classes had more education, but less expe- rience, than the A teachers. It is probable, however, that the advantage, if any, lay on the side of the B teachers. (See page 29.) Only the sixth, seventh, and eighth grades in each school participated in the experiment. The A teachers conducted their classes according to a directed-teaching plan, while the B teachers taught their classes after their own fashion. All the teachers gave the same amount of time to handwriting — namely, fifteen min- utes daily. The following written directions were placed in the hands of each directed teacher : 1. Give just fifteen minutes daily to handwriting. 2, With reference to position, teach the following points : (1) The forearm should be nearly perpendicular to the line of writing. (2) The hand should face downward, with the wrist in- clined not more than 45 degrees. (3) The hand should slide on the fourth and fifth fingers. Directed and Undirected Teaching 41 (4) The grasp of the pen should be light, with the fin- gers moderately curved. 3. With reference to movement, teach: (1) The sideward movement of the hand across the page should be free. (2) The writing movement should be a combination of the arm and finger movement. 4. Give drill exercises whenever needed to develop free- dom and fluency in movement. (See the references for these exercises.) 5. Rate each pupil's writing on both the Ayers and the Freeman Scales at the beginning of this experiment and again at the close. (The specimens will be taken by the principal of the school in each case.) 6. Lead each pupil to set up definite standards to be at- tained by the end of the school year. For this purpose util- ize the scores on the standard scales. 7. Rate each pupil's writing once a month on the Freeman Scale and keep him informed of his progress. 8. Have each pupil keep an individual record card for these monthly scores on the Freeman Scale, and plot a graph showing his own progress and the progress of the class. 9. Lead the child to analyze his own writing habits. Cen- ter attention on one important phase at a time. Distribute attention over the various phases as indicated below : First week — position and movement ; Second week — uniformity of slant ; Third week — uniformity of alignment ; Fourth week — spacing and quality of line; Fifth week — letter formation ; Sixth week — speed. At the end of the sixth week go back to position and move- ment and repeat the cycle. As soon as the standard is at- tained in any of these phases, give that phase no more spe- cial attention. 10. As far as possible, fit instruction and practice to in- dividual needs. 11. Make your instructions brief and concise in order to conserve the time for practice. 12. Lisist on all written work being done in the pupil's best quality of handwriting. 13. Keep notes on your daily work, so that you can write a full and accurate account of the experiment. 14. Make constant use of the following references : Freeman, F. N., The Teaching of Handwriting. Freeman, F. N., The Handwriting Movement. Monroe, W. S., Measuring the Results of Teaching. 42 A Comparative Study of All the teachers were urged to follow the instructions- closely, and frequent inquiry was made concerning the prog- ress of the work. No attempt was made to supervise the work of the control group of teachers, except with respect to the amount of time given to handwriting. Each teacher was left to devise her own plans from day to day. The experiment continued from October to April, a pe- riod of six months. The progress of each grade was meas- ured by the amount of gain on the Ayers Scale. The in- itial and final specimens were all rated by a board of five graders, none of whom, except the writer, knew the nature or purpose of the experiment, nor which were the directed groups or the undirected groups. The grading board, in addition to the writer, consisted of the following persons : Frank S. White, an instructor in education in the Fair- mont State Normal School and formerly a writing teacher. M. H. Cole, principal of the White School, Fairmont, W. Va. Elma Ruth Shreve, formerly a teacher of seven years' experience. T. C. Moore, an instructor in commercial subjects in the Fairmont High School, Fairmont, W. Va. Results. — Complete records were obtained from 246 pu- pils. Of this number, 120 were in the directed grades and 126 in the undirected grades. The records of 100 pupils had to be eliminated, because they were absent from one or both tests. In order to render the groups equal in numbers and approximately equal in initial achievement, six more of the poorest records were eliminated from the undirected grades, leaving 120 in each group. The computation of the coefficient of variability for each member of the grading board gave the following results : Francis Shreve .10 Elma Ruth Shreve .11 Frank S. White .12 M. H. Cole .12 T. C. Moore .12 How the directed and undirected groups compare may be seen from an examination of the following tables and graphs. In the tables A indicates the directed groups and B the undirected. Directed and JJyidirected Teaching 43 TABLE X Distribution of Scores on the Initial AND Final Rating FOR Quality Initial Rating Final Rating A B Score A B 1 68-71 9 2 1 1 65-67 8 1 1 5 62-64 6 4 1 1 59-61 5 2 2 2 56-58 10 7 2 4 53-55 3 2 9 6 50-52 17 7 5 5 47-49 8 4 19 14 44-46 13 11 9 11 41-43 12 15 18 13 38-40 8 15 14 10 35-37 7 16 18 20 32-34 8 19 7 11 29-31 3 5 11 14 26-28 3 9 1 3 23-25 1 1 20-22 Total 120 120 120 120 Median 39.3 38.5 49.3 40 P. E. 6.1 6.8 7.5 7.0 Difference .8 9.3 .93 P. E. of D .83 32 |\ 30 / /^ \ 1/ \ \ 28 // \ / / ] N^ •', 26 / / ' f / 1 -- ; / \ \ -- 22 / / \ \ \ 1 20 ,' / \ \ 18 .' / 16 ' / \\ " 7 \ 12 7 \ 10 // , \ J-J.Sft 20-25 26-31 32-37 38-43 44-49 50-55 56-61 62-67 68-73 74-78 Figure 7 — Showing the relative standing of the directed and undirected groups on the initial rating for quality, Ayers Scale. •• 44 A Comparative Study of 36 34 32 30 28 26 24 22 20 18 16 14 12 10 20-25 / 1 > \ " V,^^ / / i \^ \ -- /' / ' \ / 1 \ V \ -- \ \ -- \ \ / / \ \ \ \ N \ N \ // U.ifo \ \ \ \ \ \ -- 26-31 32-37 38-43 44-49 50-55 56-61 62-67 68-73 74-79 Figure 8 — Showing the relative standing of the directed and undirected groups on the final rating for quality, Ayers Scale. TABLE XI Distribution of Scores on the Initial and Final Tests for Rate. Initial Test A B 4 5 5 5 2 3 1 2 8 2 7 6 5 5 11 20 21 5 7 10 12 14 8 14 16 11 7 14 10 Total 120 120 Median 64.5 71 P. E. 14.1 11.6 Difference 6.5 Final Test Score 121 116-120 111-115 106-110 101-105 96-100 91-95 86-90 81-85 76-80 71-75 66-70 61-65 56-60 51-55 4e--50 -45 A 11 1 1 2 5 4 7 12 9 15 23 9 12 2 4 1 2 120 78.3 12.6 2.0 1 RQ B 2 1 3 3 6 4 5 7 13 17 23 10 9 9 5 1 2 120 76.3 11.2 Directed mid Undirected Teaching 42 39 36 33 30 27 24 21 18 ,/ / A /' \ /' \ ' / \ ^ /' w V; 9^^/fi 15 12 9 6 3 41-50 51-60 61-70 71-80 81-90 91-100 101-110 111-120 121-130 Figure 9 — Showing the relative standing of the directed and undirected groups on the initial test for speed. The score is in terms of letters per minute. 42 45 42 39 36 33 30 27 24 21 18 15 12 9 6 3 41-50 51-60 61-70 71-80 81-90 91-100 101-110 111-120 121-130 Figure 10 — Showing the relative standing of the directed and undirected groups on the final rating for speed. 46 Rate 80 76 72 68 64 60 56 52 48 44 40 36 32 28 A Comparative Study of 1 . . 1 ...X-:'^ J 1 (l k^^d ) 7 1 ^ ^ d 1 1 1 (^' DO --- fi4 ... r / ^ / y — bU ^ k' •^9 — )1 J 48 --- / AA r 1 AO V / J i^fi / 32 / 28 --- :f..- 34 38 42 46 50 54 58 62 66 70 Quality Figure 12— Showing the relative standing of the directed and undirected groups, by grades, on the final rating for speed and quality. Q\ .(T\ —(2) Ayers Standards Q f 2)" O Eirected Groups (~\ (~\ Q Undirected Groups TABLE XII Showing the Gross Gain, Percentage of Gain, and Gain in Terms OF School Years Quality Gross Directed 10 Undirected 1-5 Rate Directed 13.8 Undirected 5.3 TABLE XIII Percentage School Years 25 2.5 4 .4 21 7 2.4 1.0 Showing the Percentage of the Directed Groups Equaling or Exceeding the Median of the Undirected initial Test Final Test Gain Quality 53.3 66.6 13.3 Rate 43.3 55.0 11.7 48 A Comparative Study of Experiment 3 Silent Reading The Subjects. — The pupils of two fourth grades in the same school system acted as subjects for this experiment. 4A had an enrollment of 32, with an average age of 9.3, and 4B had an enrollment of 36, with an average age of 9.6. The Otis Group Intelligence Test, Form B, was given to both grades, with the following results : Grade Median P. E. 4A (directed) 55.3 6.3 4B (undirected) 56.0 4.6 Difference, 0.7; P. E. of difference, 1.46 26-30 31-35 36-40 41-45 46-50 51-55 56-60 61-65 66-70 Figure 13 — Showing the relative distribution of scores on the Otis Intelligence Scale. The two grades, therefore, were approximately equal in general intelligence, with a slight advantage in favor of 4B. To avoid favoring the directed group, 4A was chosen for direction and 4B for the undirected group. Teachers. — Two teachers, known as X and Y, cooperated in this experiment. They held similar positions in the same school system, but how nearly they were equal in teaching ability could not be accurately determined. On the basis of education and years of experience they ranked as follows : acher Education Years of Experience X High School 9 Y Normal 2 Teacher X had the advantage in experience, but Teacher Y had superior training. Since education has been found to be a more reliable measure of teaching efficiency than is experience, it seems probable that Teacher Y had a slight advantage over Teacher X. (See page 29.) Method of Procedure. — Teacher X directed the work of 4A according to a directed-teaching plan, while Teacher Y taught 4B, following her own plan. Both teachers gave the same amount of time to reading — namely, thirty minutes a Directed and Undirected Teaching 49 day. This time was divided into two periods of fifteen min- utes each. The same text was used in both grades. At the beginning of the experiment the writer gave Mon- roe's Standardized Silent Reading Test, Form 1, to both grades. The experiment was then continued for twelve weeks, from the middle of January to the middle of April. The teacher directing the work of 4A proceeded according to the following plan : 1. On the basis of what the Monroe test revealed the needs of the class were diagnosed. 2. The pupils of similar ability were grouped together, thus forming three groups. Group I, containing the pupils reaching or exceeding the standard in both speed and comprehension, numbering seven. Group II, containing the pupils falling below the stand- ard in both rate and comprehension, numbering eighteen. Group III, containing the pupils failing to reach the standard in comprehension, but exceeding it in rate, num- bering seven. 3. Improvised tests similar to the standard tests were given frequently, and the pupils regrouped according to their attainments. 4. Special measures were adopted designed to increase the rate of reading, particularly of Group II, such as : (1) Selecting reading matter with easy vocabulary. (2) Timed sentence tests. These tests were graded and each pupil informed of his score. Graphs were made show- ing the progress of the group. (3) Flash cards were used for phrase drills. (4) Having pupils read to see who could find answers to questions most quickly. 5. Several measures designed to increase comprehension were employed, such as: (1) Frequent word drills and studies were conducted to increase the pupil's vocabulary. Difficult words were worked out by phonics and analysis. (2) The meanings of words were studied, and the words used in oral and written sentences. (3) Reading material was selected that made a strong appeal to the pupils. 6. An attempt was made to motivate the work by utiliz- ing native interests. Pupils who enjoyed construction work were given written instructions to follow. Those who liked drawing were encouraged to illustrate poems and stories. All the pupils were encouraged to use the books and maga- 50 A Comparative Study of zines on the reading table, selecting these according to their own interests. At the end of the twelve-weeks' period Monroe's Stand- ardized Reading Test, Form I, was given a second time to both grades. Results.— After eliminating the records of all pupils not present for both tests and pairing off to equalize initial achievement, twenty-eight complete records were obtained from each grade. TABLE XIV Distributions of Scores on the Initial and Final Tests FOR Comprehension Initial Test Final Trst iA iB Score 29-30 27-28 25-26 23-24 21-22 19-20 17-18 JfA 1 2 2 2 2 1 4 AB 3 3 3 1 15-16 2 1 5 2 13-14 2 2 1 • 1 11-12 7 1 3 9 9-10 5 7 4 7-8 2 5 5 4 5-6 2 3 2 3-4 1 2 3 6 1-2 1 Total 28 28 28 28 Median 7.9 8.0 17 10.6 P. E. 2.8 2.5 4.4 4.1 Difference —0.1 6.4 P. E. of D. .70 1.13 14 -- 12 10 A -- 17-20 21-24 25-28 29-32 33-36 Figure 14 — Showing the relative attainment of the two groups on the initial test for comprehension. Directed mid Undirected Teaching 51 12 10 'd-Vl 13-16 17-20 21-24 25-28 29-32 33-36 Figure 15 — Showing the relative standing of the two grades on the final test for comprehension. TABLE XV Distribution of Scores on the Initial and Final Tests for Rate Initial Test Total 28 Median 62.7 P. E. 11.5 Diflference 6.7 P. E. of D 3.6 2 4 8 5 3 7 1 1 2 28 56 15.3 Score 126-135 116-125 106-115 96-105 86-95 76-85 66-75 56-65 46-55 36-45 26-35 16-25 6-15 Final Test 4A J,B 28 86 17.4 4.0 4.65 28 82 17.4 6-25 26-45 46-65 66-85 106-125 126-145 Figure 16 — Showing the relative attainment of the two grades on the initial test for rate. 52 A Comparative Study of 12 10 6-25 26-45 46-65 66-85 86-105 106-125 126-145 Figure 17 — Showing the relative standing of the two grades on the final test for rate. TABLE XVI Showing the Gross Gain, Percentage of Gain, and Gain in Terms OF School Years comprehension Gross Percentage School Years Directed 9.1 115 1.5 Undirected 2.6 83 .4 rate Directed 23.3 37 1.1 Undirected 26.0 46 1.2 TABLE XVII Showing the Percentage of the Directed Group Equaling or Exceeding the Median of the Undirected Initial Test Fiyial Test Gain Comprehension 48.2 89.3 41.1 Rate 64.3 46.4 —17.9 Experiment 4 Problem Solving in Arithmetic The Subjects. — The pupils of the fifth grade in two dif- ferent schools acted as subjects for this experiment. These schools were located in small towns, and had less than 5% of foreign pupils. Each grade was divided into two sec- tions, known as A and B, of approximately equal ability and initial achievement, on the basis of scores made on an in- telligence test, and an initial test with Monroe's Standard- ized Reasoning Test II, Form 1. Unfortunately, one prin- cipal failed to give the intelligence test, and in this case the grade was divided on the basis of scores made on Monroe's Standardized Reasoning Test. On the reasoning test this grade made the following scores for correct principle: Dir^ected and Uyidirected Teaching 53 ction No. of Pupils Av. Age Median p. E. A 18 11.15 18.7 4.5 B 20 11.28 18.5 4.5 On the National Intelligence Test the other fifth grade made the following record : Section A B No. of Pupils 13 12 Av. Age 11.92 12.08 ?.5 p. E. 14.9 9.4 1-4 5-8 9-12 13-16 17-20 21-24 •25-28 29-32 33-36 Figure 18 — Showing the relative distribution of the scores made on the initial test on Monroe's Standardized Reasoning Test. 41-50 Figure 19 — Showing the relative distribution of scores on the National Intelligence Test. Section A was chosen, in each case, for the directed group, and Section B for the undirected group. The experiment continued for eight weeks. The Teachers. — Four teachers, one for each section, co- operated in this experiment. All of these teachers, except one, were the regular classroom teachers, and had charge of the pupils they were accustomed to teach. How these teachers compared with respect to education and experience is shown in Table XVIII. TABLE XVIII Showing the Relative Amounts of Education and Experience of THE Teachers Cooperating in Experiment 4 Directed School Education Experience School A Elementary 8 School B College 15 Undirected Education Experience Normal 1 High School 3 54 A Comparative Study of The teacher with a college education taught Section A in one school for the purpose of the experiment. Conse- quently he taught pupils that he had not been accustomed to teach and a subject that he had not taught for ten years. Apparently these factors more than overbalanced his supe- riority due to better training and more years of experience, for the teacher with only an elementary education achieved better results with her section than the college teacher did with his section. Method of Procedure. — The directed groups and the un- directed groups gave thirty minutes daily to arithmetic, and covered the same amount of text. The undirected teachers were left to their own resources for plans and devices, while the directed teachers worked according to the following plan: 1. Diagnose the needs of your section as directed and suggested by Monroe in his Measuring the Results of Teach- ing, pages 157-174. 2. Teach a series of lessons to the section, emphasizing in turn the steps required in solving a problem, according to Monroe. Fit the instruction to the needs of the class as shown by the diagnosis. (See Monroe's Measurifig the Re- sults of Teaching, pages 160-168.) 3. Check up on the progress of the class by testing them every tenth lesson, using the Stone Scale for measuring rea- soning ability in arithmetic. Make a graph showing the standing of the class on each test. Have each pupil keep a similar graph showing the class standing and his own stand- ing. 4. Explain to the pupils the meaning of the standard scores, and lead each pupil to set up, as a goal to be reached, a definite number of points of improvement to be made dur- ing the experiment. The initial and final test were given by the principal of the school and scored by the writer. Progress was meas- ured in terms of gain on the Monroe Scale. Results.— See Tables XIX to XXII and Figures 20-23. Directed and Undirected Teaching 55 TABLE XIX Distribution OF Scores ON THE Initial and Final Tests FOR Correct Principle Initial Test Final Test 5A 5B Score 34-35 5A 2 5B 2 32-33 2 1 2 1 30-31 3 2 1 2 28-29 2 2 1 2 26-27 1 2 1 3 24-25 1 2 1 1 22-23 2 1 2 1 20-21 2 2 • 2 1 18-19 2 3 3 2 16-17 1 1 1 4 14-15 5 4 5 2 12-13 5 3 5 5 10-11 3 6 3 6 8-9 2 2 2 6-7 4-5 1 Total 31 32 31 32 Median 15 14.5 19.5 16 P. E. 5.5 5.1 5.6 5.2 Difference .5 3.5 1.4 P. E. of D 1.3 10-13 14-17 18-21 22-25 30-33 Figure 20 — Showing the relative achievement of the two groups on the initial test for correct principle, Monroe Scale. 10 2-5 6-9 10-13 14-17 18-21 22-25 26-29 30-33 34-37 38- FlGURE 21 — Showing the relative standing of the two groups on the final test for correct principle. 56 A Compay^ative Study of TABLE XX Distribution OF Scores o> r THE Initial and Final Tests FOR Correct A NSWER Initial Test Final Test 5A 5B .Score 24-25 22-23 5A 1 5B 1 20-21 2 1 2 1 18-19 3 1 2 1 16-17 4 3 2 2 14-15 4 4 5 3 12-13 6 5 3 8 10-11 6 4 9 4 8-9 4 8 4 3 6-7 1 4 2 8 4-5 2 1 2 2-3 0-1 Total 31 32 31 32 Median 9.9 9.5 13.5 11 P. E. 3.0 2.8 2.7 2.7 Difference .4 2.5 P. E. of D. .73 .68 12 r" "TTv - 10 1 / 1 ^ 8 1 1 / 1 \\ 1 ^ \ _. 6 / \ N^ \ -- 4 / . / 1 s. -- 2 j/ N. ^\ j^. . . . ; sj.f^o "^ " ^ irr:^^ ^ 0-3 4-7 8-11 12-15 16-19 20-23 24-27 2^-31 Figure 22 — Showing the relative initial achievement of both groups on the initial test for correct answer. 0-3 4-7 8-11 12-15 16-19 20-23 24-27 28-31 Figure 23 — Showing the relative standing of the two groups on the final test for correct answer. Dir^ected and Undirected Teaching 57 TABLE XXI Showing the Gross Gain, Percentage of Gain, and Gain in Terms of School Years CORRECT principle Gross Percentage School Years Directed 4.5 30 .6 Undirected 1.5 10 .2 correct answers Directed 3.6 36 .8 Undirected 1.5 16 .3 TABLE XXII Showing the Percentage of the Directed Groups Equaling or Exceeding the Median of the Undirected Initial Test Final Test Gain Correct Principle 51.6 58 6.4 Correct Answer 53.9 74.2 20.3 Experiment 5 Appreciation of Poetry The Subjects. — The subjects for this experiment con- sisted of all the pupils of three seventh grades in two schools located in the residence sections of a city of about 18,000 inhabitants. The pupils were all of native American par- entage, and many of them came from homes of the business and professional men of the city. They were divided into two groups, known as A and B, approximately equal in mental ability and initial achievement, on the basis of the scores made on a group test for mental ability and on a special test designed to measure the ability to appreciate poetry. Terman's Group Test for Mental Ability, Form A. was used to measure the mental ability of the pupils, and a special test designed by the writer was used to measure their ability to appreciate poetry. This test is fully de- scribed in the following pages. On the basis of these tests the A groups were selected for the directed group and the B groups for the undirected group. The directed group had nearly twice as many pupils as the undirected group, but the directed group was divided into two sections and taught by two teachers. For the scores on the Terman Test for Men- tal Ability, see Table XXIII and Figure 24 ; for the scores on the initial test for appreciation of poetry, see Table XXIV and Figure 25. 58 A Comparative Study of TABLE XXIII Distribution of Scores ON THE TERMAN Ability Test for Mental Score Directed Undirected 161-170 1 151-160 1 2 141-150 3 1 131-140 3 1 121-130 4 1 111-120 6 4 101-110 6 4 91-100 5 5 81-90 4 4 71-80 7 5 61-70 7 2 51-60 6 3 41-50 2 1 31-40 1 Total _ 56 33 Median _ 93 94 P. E. _ 21.1 14.1 Difference _— 1 _ 3.73 P. E. of D 9 -- 8 7 , -- 6 / 1 \ -- 5 1 \ \ 4 / / \ \ 3 / / / / \ " 2 / /" \__ ~ '^ \ 1 y / l^/.2-^ \\ 0' (1 \. 31-45 46-60 61-75 76-90 91-105 106-120 121-135 136-150 151-165 Figure 24 — Showing the relative distribution of scores on the Terman Test for men- tal ability. The Teachers. — Three teachers cooperated in this experi- ment — two for the directed group and one for the undi- rected group, the directed group being approximately twice as large as the undirected group. With respect to educa- tion and experience in teaching, the directed teachers were approximately equal to the undirected teachers, as may be seen from the following data: Directed and JJyidirected Teaching 59 Teacher • Education Years of Experience Directed Normal 18 Directed Normal 5 Undirected Normal 11 Method of Procedure. — The following scale for measuring the appreciation of poetry was devised and tentative stand- ards established by giving the test in five different schools in the seventh and eighth grades. Records were obtained from 129 pupils in the seventh grade and 127 in the eighth grade. The seventh-grade pupils made a median score of 22, and the eighth-grade pupils made a median score of 28 — a difference of 6 points. The poem used in the test and the directions for giving and scoring the tests were typewritten on separate sheets of paper. Score A Scale for Measuring the Appreciation of Poetry Name School Grade Age Place a copy of "A Noble Nature" in the hands of each pupil. Directions to the Pupils Read this poem carefully, then do as directed below. Be careful to do exactly as directed. 1. A poem may have as its main content any or all of the following: (1) An important idea. (Thought.) (2) Humor. (Fun.) (3) Sense experience. (Suggestions of sight, sound, odor, or taste.) Place a cross before the above statement that best expresses the nature of this poem — e. g., if this poem contains an important idea, place a cross before statement No. (1) ; if it is a humorous poem, be- fore statement No. (2) ; etc. 2. Select from this poem two lines that best express the meaning. Write these lines in the space below. 3. In giving effective expression to their thoughts and feelings, poets commonly employ the following: (1) Images or mental pictures. (2) Rhythm and rhyme. (3) Well-chosen words and phrases. (4) Interesting stories. (5) 'Strong and appealing characters. (6) Figures of speech. Place a cross before any of the above means or devices that the author has used very effectively in this poem. More than one may be marked. 4. A poem may arouse one or more of the following emotions: (1) Joy or gladness. (2) Faith or trust. (3) Aspiration. (4) Admiration. (5) Patriotism. 60 ■ A Comparative Study of (6) Sadness or grief. (7) Love or sympathy. (8) Awe or mystei'y. (9) Hopeless despair. Place a cross before any of the above statements that tell how this poem makes you feel. More than one may be marked. 5. A poem may be used in any of the following ways: (1) Read for pleasure. (2) Read to modify one's own conduct. (3) Read to others for their instruction or pleasure. (4) As a basis for language and grammar lessons. (5) Memorized and quoted to others. Place a cross before any of the above statements that tell how this poem should be used. More than one may be marked. A definite set of directions for administering and scoring the tests was written out and applied in every instance. The scoring was done on the basis of 50 points for a perfect score. These points were distributed arbitrarily as follows : v'o. of Exercise PoitU 's Allowed 1 15 2 10 3 10 4 10 5 5 Total pc lints- _ 50 The poem used for these tests was "A Noble Nature." Each of the three grades acting as subjects for this ex- periment was tested on the preceding scale. Then the di- rected teachers taught their groups, according to a definite plan for the teaching of poetry, three lessons per week for a period of ten weeks. The class periods were thirty min- utes in length. Meanwhile the undirected teacher gave the same amount of time to the teaching of poetry, following her own plans and devices. The same poems were taught to both groups. The directed teacher gave emphasis to what constitutes excellence in poetry and to teaching the pupils how to study poetry. To this end definite assignments were made and specific directions for study given. In general, an attempt was made to lead the pupils to see what constitutes excel- lence in poetry. The poems studied, fifteen in all, were selected from the Elson Grammar School Literature, Book I, which was the text then in use in the city schools. At the end of the ten weeks both groups were tested a second time on the scale previously described, using the same poem as a basis for the test, and the progress of each group measured in terms of gain over the previous record. Directed ayid Undirected Teaching 61 It should be noted that the pupils were not tested on the poems studied in the course of the experiment, but on a scale designed to measure their ability to appreciate poetry. The gain, therefore, represents not an increase in the spe- cific knowledge of the poems studied, but an increase in the ability to appreciate poetry. Results. — For a summary of the results, see Table XXIV and Figures 25 and 26. TABLE XXIT Distribution of Scores ON THE Initial and Final Test FOR Appreciation Initial Test Final Test 7A 7B Score 7A 7B 1 40-41 1 1 1 1 37-39 3 1 3 1 34-36 9 3 4 2 31-33 10 3 6 4 28-30 10 4 8 6 25-27 5 4 10 5 22-24 4 5 6 4 19-21 4 4 4 3 16-18 3 3 3 2 13-15 3 2 3 2 10-12 2 2 2 1 7-9 2 1 2 1 4-6 2 1 1-3 1 Total .__ 56 33 56 33 Median .__ 23.5 23.5 29.5 24.7 P. E. .__ 6.3 5.7 ' 5.6 5.5 Difference .__ 4.8 P. E. of D .__ 1.3 1.21 18 13-18 19-24 25-30 43-48 Figure 25 — Showing the relative achievement of the two groups on the initial test for the appreciation of poetry. 62 A Comparative Study of 1-6 7-12 13-18 19--4 25-30 31-36 37-42 43-48 Figure 26 — Showing the relative standing of the two groups on the final test for the appreciation of poetry. TABLE XXV Showing the Gross Gain, Percentage of Gain, and Gain in Terms of School Years Gross Percentage School Years Directed 6 26 1.0 Undirected 1.2 5 .2 The percentage of the directed groups equaling or exceed- ing the median of the undirected groups was as follows : Initial Test Final Gain 50 67.9 17.9 Experiment 6 History in the High School The Subjects. — The subjects for this experiment con- sisted of two classes in history — one in American history and one in modern European history — in Fairmont High School, Fairmont, W. Va. The American history class, numbering 59, was taught by Miss Elizabeth Koletka, and the modern European class, numbering 30, by Mr. H. Y. Clark. Both of these teachers were college graduates, with some special training for the teaching of history. Each class was divided into two sections, known as A and B, approximately equal in size, in mental ability, and in initial achievement in history. The division was made on the basis of the scores made on Terman's Group Test of Mental Ability, Form A, and a series of special tests in history, devised in each case by the teacher in charge of the class. (For the scores on Terman's Group Test of Mental Ability, see Table XXVI and Figure 27.) The spe- cial tests devised by the teachers covered the subject-matter Directed mid Undirected Teaching 63 that the pupils had studied the first half of the school year, and were somewhat technical. Three tests of fifteen min- utes each were given to each class on three successive days, and the average score taken as the true score. The ques- tions for these tests were framed so that each question re- quired a definite answer. Consequently they could be graded with a fair degree of accuracy. Moreover, each teacher made out a definite grading plan, indicating what answers would receive full credit and what only partial credit, or no credit. The questions were weighted by allow- ing a certain number of points for each question answered correctly, the number of points being determined by the difficulty of the question as estimated by the teacher's judg- ment. The 100-point scale was used. In order to make the grading still more accurate, the papers were scored by the writer, and the average of his grade and the teacher's grade taken as the true score for each pupil. (For the scores on these special, initial tests in history, see Table XXVII and Figure 28.) Method of Procedure. — The experiment continued for nine weeks, the last half of the last semester of the school year of 1920-21. Both sections of each class were taught by the same teacher. In each case the teacher directed the work of Section A, following a directed-teaching plan, and taught Section B by the usual plan. The periods were sixty minutes in length. With Section A the period was divided, thirty minutes being allotted to the recitation and thirty minutes to supervised study. Section B had forty-five min- utes for the recitation and fifteen minutes for study with- out supervision. An attempt was made to teach Section A how to study history. To this end emphasis was placed upon factors, such as: Finding the main points ; Making outlines ; Underscoring and making marginal notations ; Exercising judgment in the selection and organization of material and in the interpretation of historical data ; Using the reference books. Minimum, average, and maximum assignments were made, thus providing for individual differences in ability to do work. The progress of the class was checked fre- quently by giving short-time tests. The test papers were graded carefully, and each pupil informed of his standing. Graphs were made showing the standing of the class on each test, and each pupil was encouraged to compare his standing with the class standing. 64 A Comparative Study of During the study period the teacher passed quietly from pupil to pupil, giving help according to individual needs. A pupil desiring help was required to raise his hand. Help was given largely in the form of questions and suggestions. At the close of the experiment another series of tests was given on the subject-matter that had been covered during the experiment. This series of tests was administered and scored in the same manner as the initial series. In the sec- ond series the pupil's name or number was written on the back of the paper, so that the teacher did not know whose paper he was grading, except as he recognized the hand- writing. The writer was not personally acquainted with any of the pupils; hence, his grading was free from any personal bias. All the tests, in both the initial and the final series, were made long enough to keep every pupil busy for full fifteen minutes, and so difficult that no pupil could make a perfect score, but, at the same time, so simple that the slowest pupil could answer a few of the questions. In addition to the tests described in the preceding par- agraphs, Harlan's Standard Test in American history was given at the beginning of the experiment and again at the close to the two sections studying American history. No standard test was available for the sections studying mod- ern European history. Results. — Complete records were obtained from thirty pupils in American history, fifteen in each section, and fifty-five in modern European history, twenty-five in Sec- tion A and thirty in Section B. Four records had to be eliminated from Section A because of absence from tests. TABLE XXVI Distribution of Scores on Terman's Group Test of Mental Ability 5(0>e Sec. A Sec B 201-210 1 191-200 1 181-190 1 171-180 1 1 161-170 2 2 151-160 2 3 141-150 2 4 131-140 5 6 121-130 7 5 111-120 4 4 101-110 3 8 91-100 5 3 81-90 5 2 71-80 1 2 61-70 1 2 51-60 1 1 Total 40 45 Median 121 122 P. E. 21.1 22.2 Difference — i P. E. of D 4.69 Directed and Undirected Teaching 65 12 12 -50 51-70 71-93 91-110 111-130 131-150 151-170 171-190 Figure 27 — Showing the relative attainment of the two groups on Terman's Group Test of Mental Ability. TABLE XXVII Distribution of Scores on the Special Initial and Final IN History Tests Initial Test Final Test Sec. A Sec. B Score 96-100 91-95 Sec. A 2 4 Sec.B 1 4 86-90 6 2 1 2 81-85 11 4 4 4 76-80 3 7 6 4 71-75 2 2 8 12 66-70 4 15 2 2 61-65 3 3 6 4 56-60 1 5 4 6 51-55 1 3 4 4 46-50 2 3 1 2 41-45 2 1 36-40 1 1 1 31-35 . 26-30 Total 40 45 40 45 Median 66 67.5 82.4 10.1 68.5 P. E. 9.1 9.0 7.6 Difference —1.5 13.9 P. E. of D 1.97 1.95 66 A Comparative Study of 21-30 31-40 41-50 51-60 61-70 71-SO 81-90 91-100 Figure 28 — Showing the relative achievement of the two groups on the special initial tests in history. 31-40 41-50 51-60 61-70 71-80 81-90 91-100 Figure 29 — Showing the relative standing of the two groups on the special final tests in history. TABLE XXVIII Distribution of the Scores on Harlan's Standard Test in American History Initial Test Final Test Sec. A Sec. B Score 96-100 Sec. A 1 Sec. B 91-95 2 1 86-90 3 1 2 81-85 4 4 5 1 76-80 5 3 ^ 2 5 71-75 3 8 2 7 66-70 2 4 5 3 61-65 2 3 4 9 56-60 1 4 2 2 51-55 1 2 2 1 46-50 1 1 41-45 1 36-40 1 31-35 26-30 Total 25 30 25 30 Median • 62.5 66 78.5 72.3 P. E. 7.0 5.6 8.6 6.9 DifTerpncf" „_ ^ —3.5 6.2 2.13 P. E. of D 1.73 Directed and Undirected Teaching 67 21-30 31-40 41-50 51-60 61-70 71-80 81-91 91-100 Figure 30 — Showing the relative initial achievement of the tv7o groups on Harlan's Test of Information in American History. 21-30 31-40 41-50 51-60 61-70 71-80 81-90 91-100 Figure 31 — Showing the relative standing of the two groups on the final test with Harlan's Test of Information in American History. TABLE XXIX Showing the Gross Gain, Percentage of Gain, and Gain in Terms of School Years Special Tests in History School Gross Percentage Years Section A (directed) 16.4 25 Section B (undirected) 1.0 2 Harlan's Standard Test Section A (one class only) 16.0 26 .5 Section B (one class only) 6.3 10 .2 TABLE XXX Showing the Percentage of Section A Equaling or Exceeding THE Median of Section B Initial Test Final Test Gain Special Tests 47.5 66.7 19.2 Harlan's Test 36.0 68.0 32.0 Experiment 7 History in the Grades The Subjects. — The subjects for this experiment con- sisted of two sections of the seventh grade of a school lo- cated in a city of about 20,000 inhabitants. Approximately 68 A Comparative Study of 5% of the pupils of this grade were of foreign parentage. These pupils were equally divided between the two sections. The Teacher: Both sections were taught by Miss Ruby Shaffer, a normal-school graduate, with three years' expe- rience in teaching. Miss Shaffer was fully aware of the nature of the experiment, and made an honest effort to keep all the important factors as nearly uniform as possi- ble, making only such differences as were required by the specific plan for directing the work of Section A. Method of Procedure. — The pupils of the seventh grade were divided into two sections, known as A and B, approxi- mately equal in numbers, in age and mental ability, and in initial achievement in history. The division was made on the basis of the scores made on the Haggerty Intelligence Test, on Van Wagenen's Standard Test in American His- tory, and on a special test devised by the teacher, covering the work that had been done since the beginning of the school year. (For the scores on these tests, see Tables XXXI to XXXV and Figures 32, 33, and 35.) The experiment continued from the first of March to the middle of May, 1921 — a period of ten weeks. Both sec- tions gave the same amount of time to history and covered the same subject-matter, both in kind and amount. But Section A worked under a directed-teaching plan, while Section B worked under the customary plan. At the close of the experiment. Van Wagenen's Standard Test, Form A, was repeated. The same form was used for both initial and final tests. The information scale, the character judgment scale, and the thought scale were all used, and the average of the three scores taken as the final score. These tests were administered and scored by the principal of the school. A second special test, covering the work done in the ten weeks, was devised and administered by the teacher. The papers of this test, as well as the in- itial test, were all graded by the teacher and the principal, and the average of the two grades taken as the true score. Results. — The results are summarized in Tables XXXII to XXXV and Figures 32 to 36. Complete records were obtained from fifty-one pupils — twenty-five in Section A and twenty-six in Section B. Directed and Undirected Teaching 69 TABLE XXXI Distribution OF Score 3 ON THE Haggerty Intelligence Test Score Sec. A Sec. B 136-140 1 131-135 ' 126-130 1 3 121-125 1 2 116-120 1 2 111-115 2 3 106-110 4 2 101-105 2 5 96-100 5 2 91-95 2 1 1 86-90 1 81-85 1 1 76-80 2 2 71-75 1 66-70 2 2 61-65 56-60 Total _ 25 _ 99.5 _ 11.1 _ —6.5 26 IVTorli a Y\ 106 iVitrUldil P F! 12.7 P F nf D _ 3.34 i, JZi. \J± xy. — — 7 u I 7 6 / 6 5 / \ / 1 \ / 1 \ \ \ 5 4 / / 1 \ 3 -^ i 1 / / / / \ 3 \ 2 2 / " N / ^ / s. \ / \ / \ \ 1 / n /- ■J \36Yo \ 1 > 51-60 61-70 71-80 81-90 91-100 101-110 111-120 121-130 131- FlGURE 32— Showing the relative attainment of the two sections on the Haggerty Intelligence Test. The average age of the two sections was : Section A 13-28 Section B 1-5.19 On the basis of general intelligence there was a slight dif- ference in favor of Section B. 70 A Comparative Study of TABLE XXXII Distribution of Scores on Van Wagenen's Standard Test in American History Initial Test Final Test Sec. A Sec. B Score 19-19.9 Sec. A 1 Sec. B 1 1 18-18.9 2 1 1 17-17.9 2 2 1 2 16-16.9 1 3 3 2 15-15.9 2 1 2 2 14-14.9 3 2 1 1 13-13.9 3 2 2 12-12.9 1 3 3 2 11-11.9 2 4 5 3 10-10.9 2 3 9-9.9 3 3 3 2 8-8.9 1 1 4 2 7-7.9 2 1 1 1 6-6.9 1 1 1 1 1 5-5.9 4-4.9 3-3.9 1 Total 25 26 25 26 Median 10.7 11 13.5 12.3 P. E. 2.5 2.6 2.6 2.2 Difference —0.3 1.2 P. E. of D .71 .67 13-14 15-16 17-18 19-20 Figure 33 — Showing the relative initial achievement of the two sections on Van Wagenen's Standard Test in History. 11-12 13-14 15-16 17-18 19-20 4 2 21- Figure 34 — Showing the standing of the two sections on the final test with Van Wagenen's Standard Test in American History. Directed and Undirected Teaching 71 TABLE XXXIII Distribution of Scores on the Special Initial and Final Tests IN History Initial Test Final Test Sec. A Sec. B Score 96-100 Sec. A 3 Sec. 2 1 91-95 4 2 6 5 86-90 5 2 2 4 81-85 3 4 5 5 76-80 2 5 2 2 71-75 1 2 2 4 66-70 2 2 3 61-65 3 • 1 1 56-60 3 1 2 51-55 1 1 1 46-50 2 2 41-45 1 1 1 36-40 31-35 26-30 21-26 1 Total 25 26 25 26 Median 78.5 78 85.2 76 P. E. 9.7 9.9 10.9 10. Difference .5 9.2 3.02 P. E. of D 2.74 21-30 41-50 51-60 61-70 81-90 91-100 Figure 35 — Showing the relative initial achievement of the two sections on the spe- cial initial test in history. 10 21-30 31-40 41-50 51-60 61-70 71-80 81-90 91-100 10 8 6 4 2 - Figure 36 — Showing the standing of the two sections on the special, final test in history. 72 A Compaf'ative Study of TABLE XXXIV Showing the Gross Gain, Percentage of Gain, and Gain in Terms of School Years Van Wagenen's Standard Test Gross Percentage School Years Section A (directed) 2.8 26 1.1 Section B (undirected) 1.3 12 .5 Special Tests Section A 6.7 8 , Section B —2.0 —3 TABLE XXXV Showing the Percentage of Section A Equaling or Exceeding THE Median of Section B Initial Test Final Test Gain Van Wagenen's Test 44 60 16 Special Test 52 68 16 Experiment 8 Biology The Subjects. — Two sections of a class in biology in the Fairmont High School, Fairmont, W. Va., served as sub- jects for this experiment. The class, numbering thirty- four, was a fairly representative group, coming from va- rious sections of the city and representing widely different intellectual levels. The Teacher. — Both sections were taught by Miss Frances Frost, who was a college graduate, with some spe- cial training for the teaching of biology. Miss Frost was interested in determining which was the most effective method of teaching a class, and made an effort to give each method a fair trial. Method of Procedure. — The class was divided into two sections, approximately equal in numbers, in mental ability, and in initial achievement in biology. The division was made on the basis of the scores made on Terman's Group Test of Mental Ability, Form A, and on a series of special tests devised by the teacher, covering the work that had been done in biology since the beginning of the school year. (For the scores on these tests, see Tables XXXVI to XXXVIII and Figures 37 and 38. The experiment continued for nine weeks, from March 14 to May 16, 1921. During this time Section A was taught according to a directed-teaching plan and Section B by the usual plan. Section A had thirty minutes for the recita- tion and thirty minutes for supervised study, while Section B had fifty minutes for recitation and prepared their les- sons at home or in study periods without any supervision. Directed mid Undirected Teaching 73 Both sections had the same field work and covered the same subject-matter. At the close of the experiment another series of three spe- cial tests was devised and administered by the teacher. These tests were fifteen minutes in len^h. The average made on the three tests was taken as the true score. The papers were graded by the teacher and by the writer, and the average of the ratings taken as the pupil's final score. This was true both of the initial and final series of special tests. Results. — Twenty-seven complete records were obtained — fourteen in Section A and thirteen in Section B. TABLE XXXVI Distribution of the Scores on Term an 's Group Test of Mental Ability Score Sec. A Sec. B 171-180 1 161-170 151-160 1 141-150 1 131-140 1 2 121-130 2 3 111-120 2 1 101-110 2 2 91-100 2 1 81-90 2 1 71-80 1 61-70 1 51-60 1 41-50 31-40 Total 14 13 Median HI 113.5 P. E. 16 21.4 Difference — 2.5 P. E. of D '^•Sl 5 5 31-50 51-70 71-90 91-110 111-130 131-150 151-170 171-190 Figure 37— Showing the relative attainment of the two groups on Terman's Group Test of Mental Ability. 74 A Comparative Study of TABLE XXXVII Distribution of Scores on the Initial and Final Tests in Biology Initial Test Final Test Sec. A Sec. B Score Sec. A Sec. B 96-100 1 1 1 91-95 1 1 1 86-90 3 1 1 2 81-85 4 1 2 2 76-80 3 2 1 71-75 1 1 2 3 66-70 2 61-65 1 2 2 2 56-60 1 1 51-55 1 2 46-50 1 1 41-45 1 1 36-40 1 31-35 26-30 Total 14 13 14 13 Median 68.5 68.5 83.5 69.8 P. E. 12.4 10.7 5.8 9.9 Difference 0. 13.7 P. E. of D 4.44 3.15 Figure 38 in biology. 31-40 41-50 51-60 61-70 71-80 81-90 91-100 Showing the relative initial achievement of the two sections on the tests 21-30 31-40 41-50 51-60 61-70 71-80 81-90 91-100 Figure 39 — Showing the standing of the two sections on the final tests in biology. TABLE XXXVIII Showing the Gross Gain and Percentage of Gain Gross Section A (directed) 15 Section B (undirected) 1.3 Percentage 22 2 The percentage of Section A equaling or exceeding the median of Section B was : Initial Test 50 Final Test 92.8 Gain 42.8 CHAPTER IV SUMMARY AND CONCLUSIONS Summary. — Complete records were obtained from 653 pupils from seven different schools, one of which was a high school. The experiments comprised all the grades from the fourth to the eighth and the high school. The results are brought together in summary form in Tables XXXIX to XLII. In Table XXXIX the number of pupils in each instance is the total number of pupils participating in the experiment. They are about equally divided be- tween the directed and undirected sections or groups. A 7ninus difference in the score indicates a difference in favor of the undirected section. In a few instances the compari- sons are made on both standard and special tests. In such cases the result of both tests are summarized in the tables, "St." indicating the standard test and **Sp." the special test. In order to determine the effect of directed teaching upon the brighter and slower pupils, a comparison is made of the gains of the upper and lower quartiles. (See Table XLII, pages 130, 131.) TABLE XXXIX Showing the Initial and Final Scores and the Gross Gain Subject No. Pupils Initial Final Gain Dif. Arithmetic, Ats. Directed 21 20.3 30.4 10,1 Undirected 20 21.0 24.0 3.0 7.1 Arithmetic, Rts. Directed 21 8.5 16.5 8.0 Undirected 20 13.3 15.0 1.7 6.3 Handwriting, Q. Directed 120 39.3 49.3 10.0 Undirected 120 38.5 40.0 1.5 8.5 Handwriting, R. Directed 120 64.5 78.3 13.8 Undirected 120 71.0 76.3 5.3 8.5 Reading, Com. Directed 28 7.9 17.0 9.1 Undirected 28 8.0 10.6 2.6 6.5 Reading, Rate Directed 28 62.7 86.0 23.3 Undirected 28 56.0 82.0 26.0 —2.7 Arithmetic, Cor. P. Directed 31 15.0 19.5 4.5 Undirected 32 14.5 16.0 1.5 3.0 76 A Comparative Study of Subject No. Pupils Initial Final Gain Dif. Arithmetic, Cor. A. Directed 31 9.9 13.5 3.6 Undirected 32 9.5 11.0 1.5 2.1 History, H. S., Sp. Directed 40 66.0 82.4 16.4 Undirected 45 67.5 68.5 1.0 15.4 History, Gr., St. Directed 25 10.7 13.5 2.8 Undirected 26 11.0 12.3 1.3 1.5 History, Gr., Sp. Directed 25 78:5 85.2 6.7 Undirected 26 78.0 76.0 —2.0 8.7 Poetry, H. S., Sp. Directed 56 23.5 29.5 6.0 Undirected 33 23,5 24.7 1.2 4.8 Biology. Directed 14 68.5 83.5 15.0 Undirected 13 68.5 69.8 1.3 13.7 TABLE XL Showing the Percentage of Gain and the Gain in Terms of School Years Subject Per. Gain Dif. School Years Dif. Arithmetic, Ats. Directed 49 1.2 Undirected 14 35 .45 .75 Arithmetic, Rts. Directed 94 1.75 Undirected 13 81 .58 1.17 Handwriting, Qual. Directed 25 2.5 Undirected 4 21 .38 2.12 Handwriting, Rate. Directed 21 2.4 Undirected 7 14 1.0 1.4 Reading, Com. Directed 115 1.5 Undirected 33 82 .4 1.1 Reading, Rate. Directed 37 1.1 Undirected 46 —9 1.2 —0.1 Arithmetic, Cor. P. Directed 30 .6 Undirected 10 20 .2 .4 Arithmetic, Cor. A. Directed 36 .8 Undirected 16 20 .3 .5 Directed and Undirected Teaching Subject Per. Gain Dif. School Years His ory, Grade, St. Directed 26 1.1 Undirected 12 14 .5 History, Grade, Sp. Directed 8 Undirected — 3 11 History, H. S., Sp. Directed 25 Undirected 2 23 Poetry. Directed 26 1.0 Undirected 5 21 .2 Biology. Directed 22 Undirected 2 20 77 Dif. .6 TABLE XLI Showing the Percentage of the Directed Groups Equaling or Exceeding the Medians of the Undirected Subject Initial T. Va. From Final T. Va. From Gain Arithmetic, Ats. 42.8 —7.2 71.4 21.4 28.6 Arithmetic, Rts. 33.3 —16.7 57.1 7.1 23.8 Handwriting, Qual. 53.3 3.3 66.6 16.6 13.3 Handwriting, Rate 43.3 —6.7 55.0 5.0 11.7 Reading, Com. 48.2 —1.8 89.3 39.3 41.1 Reading, Rate 64.3 14.3 46.4 —3.6 —17.9 Arithmetic, Cor. P. 51.6 1.6 58.0 8.0 6.4 Arithmetic, Cor. A. 53.9 3.9 74.2 24.2 20.3 Poetry 50.0 0.0 67.9 17.9 17.9 History, H. S. Sp. 47.5 —2.5 66.7 16.7 19.2 History, Gr. St. 44.0 —6.0 60.0 10.0 16.0 History, Gr. Sp. 52.0 2.0 68.0 18.0 16.0 Biology 50.0 0.0 92.8 42.8 42.8 Note. — According to Starch, a variation of less than 10 above or below 50% is not significant. TABLE XLII Showing the Gross Gain and Percentage of Gain of the Upper AND Lower Quartiles Upper Quartile Lower Quartile Subject Gross Per. Gross Per. Arithmetic, Attempts. Directed 7.8 30 5.0 38 Undirected 2.6 10 4.6 35 Arithmetic, Rights. Directed 6.0 23 5.8 208 Undirected 3.8 20 5.4 108 Handwriting, Quality. Directed 3.1 5 12.7 50 Undirected —0.6 —1 4.3 15 78 A Comparative Study of Subject Gross Handwriting, Rate. Directed —13.9 Undirected —11.0 Reading, Comprehension. Directed 12.6 Undirected 6.3 Reading, Rate. Directed 30.0 Undirected 25.0 Arithmetic, Correct P. Directed 3.7 Undirected — 1.9 Arithmetic, Correct A. Directed 2.8 Undirected 1.2 Poetry. Directed — 5.0 Undirected —2.9 History, H. S., Special. Directed 11.4 Undirected — 6.0 History, Grade, Standard. Directed 1.5 Undirected .5 History, Grade, Special. Directed 4.4 Undirected —3.0 Biology. Directed 3.3 Undirected — 7.0 Gross Per. —1 —1 33.2 15.8 72 33 100 55 6.8 4.2 212 221 38 31 20.0 26.0 54 123 15 —7 2.8 3.4 35 52 11 8 3.5 3.0 58 60 —19 —10 5.0 7.9 55 78 14 —7 17.4 10.3 39 21 10 3 5.0 —0.1 55 —00.5 5 —3 12.5 .8 23 2 4 —8 13.0 —70 31 —15 Conclusions. — In the foregoing experiments there are so many variable factors that could not be controlled rigidly, under ordinary school conditions, that the psychology of the results obtained remains somewhat obscure. For edu- cational practice, however, the outstanding fact is that, un- der the directed-teaching plan applied in these experiments, the directed group made more gain than the undirected, with one exception — namely, in rate of reading. On the other hand, since there are so many variable factors in- volved, it would be hazardous to draw any sweeping con- clusions. The following factual conclusions are obvious : 1. In six of the eight experiments there is a significant difference in favor of the directed groups, the difference in the medians being from four to ten times the probable error of the difference. 2. In two of the experiments — history in the seventh Directed and Undirected Teaching 79 grade and problem solving in arithmetic — the difference is small, but in favor of the directed groups. In these experi- ments the difference is from 1.8 to 3.7 times the probable error of the difference. 3. In gain in terms of school years there is a decided dif- ference in favor of the directed groups, with the exception of the gain in rate of reading. (See Table XL, page 127.) 4. On the basis of the percentage of the directed group equaling or exceeding the median of the undirected, there is a significant gain in favor of the directed groups in eleven of the thirteen comparisons. In one instance the gain is small, and in another there is a significant difference in fa- vor of the undirected group. (See Table XLI, page 129.) 5. In the directed groups the lower quartiles make more gross gain than the upper quartiles in nine out of thirteen comparisons, and a greater percentage of gain in eleven out of thirteen comparisons. In the undirected groups the lower quartiles make more gross gain in eleven out of thir- teen comparisons, and a greater percentage of gain in twelve out of thirteen comparisons. (See Table XLII.) From these figures it is apparent that directed teaching tends to help the brighter pupils more than the slower ones, but the difference is so slight that it is practically negligible. BIBLIOGRAPHY No attempt is made to present a complete bibliography. Only those studies bearing most directly on the present problem are listed. Atchison, A. E.: Helping Teachers in Service. Elementary School Journal, 20: 65-67, 1919. Allen, I. M.: Experiments in Supervised Study. School Review, 25: 398-411, 1917. Anderson, J. C. : The Use of the Woody Scales for Diagnostic Pur- poses. Elementary School Journal, 18: 770-781, 1918. Arps, G. F. : Attitude as a Determinant in Spelling Efficiency in Imme- diate and Delayed Recall. Journal of Educational Psychology, 6: 409-418, 1915. Ashbaugh, E. J.: The Measurement of Language. Journal of Edu- cational Research, 4: 32-39, 1921. Ballou, F. W. : Scale for the Measurement of English Composition. Harvard-Newton Bulletin, No. 11. Batson, W. H.: The Acquisition of Skill. Breed, F. S. : The Measured Results of Supervised Study. School Re- view, 27: 186-204 and 262-284, 1919. Breslich, E. R. : Teaching High School Pupils How to Study. School Review, 20: 505-515, 1912. Brown, J. C. : An Investigation of the Value of Drill Work in the Fundamentals of Arithmetic. Journal of Educational Psychol- ogy, 3: 485-492 and 561-570, 1912. Bryan, W. L., and Harter, N.: Studies in the Physiology and Psychol- ogy of the Telegraphic Language. Psychology Review, 4: 27-53, 1897. and 6: 348-375, 1899. Buckingham, B. R.: Spelling Ability — Its Measure and Distribution. Teachers' College Contributions to Education, No. 59, 1913. Burr, A. W.: Directed Study. School Review, 27: 90-100, 1919. Chapman, J. C, and Rush, G. P.: The Measurement of Classroom Products, 1917. Charters, W. H. : Constructing a Language and Grammar Scale. Journal of Educational Research, 1: 249-257, 1918. Cole, T. R.: One Year of Supervised Study. School Review, 25: 331- 335, 1917. Courtis, S. A.: Measuring the Effects of Supervision in Geography. School and Society, 10: 61-70, 1919. Cummins, R. A. : Improvement and Distribution of Practice. Teach- ers' College Contributions to Education, No. 97, 1919. Dearborn, W. F. : The Psychology of Reading, Archives of Psychol- ogy, No. 4. Downing, E. R. : Supervised Study and the Science Laboratory. School Review, 25: 646-651, 1917. Dunn, Grace A. : The Value of Supervised Study. Teachers' College Record, 18: 430-437, 1917. Earhart, L. B. : Teaching Children to Study, 1909. Edwards, A. S. : The Fundamental Principles of Learning and Study, 1920. Erickson, J. E.: Result of Supervised Study in the Houghton (Mich.) High School. School Review, 24: 752-758, 1916. Evans, J. W., and Knoche, F.: The Effect of Special Drill in Arith- metic as Measured by the Woody and Courtis Arithmetic Test. Journal of Educational Psychology, 10: 263-276, 1919. Directed and Undirected Teaching 81 Freeman, F. N.: How Children Learn, 1917. : The Handwriting Movement. Supplementary Ed- ucational Monographs, No. 9, 1918. : The Teaching of Handwriting, 1914. : An Analytic Scale for Handwriting. Elementary School Journal, 15: 432-441, 1915. Fulton, J. M.: An Experiment in Teaching Spelling. Pedagogical Seminary, 2: 287-289, 1914. Gates, A. I. : Recitation as a Factor in Memorizing. Archives of Psychology, No. 40, 1917. Gray, W. S.: Studies of Elementary School Reading Through Stand- ardized Tests. Supplementary Educational Monographs, No. 1. : Principles of Method in Teaching Reading as De- rived from Scientific Investigation. Eighteenth Year Book, Part II, National Society for the Study of Education. : The Technique of Supervising High School Teach- ing. School Review, 27: 512-522, 1919. : Diagnostic and Remedial Types in Reading. Jour- nal of Educational Research, 4: 1-15, 1921. Gray, C. T.: Types of Reading Ability as Exhibited Through Tests and Laboratory Experiments. Supplementary Educational Mon- ograph, No. 5. Haggerty, M. E.: The Laws of Learning. Psychological Review, 20: 411-422. Hahn, W. H., and Thorndike, E. L. : Some Results in Addition Under School Conditions. Journal of Educational Psychology, 5: 65- 84, 1914. Hall-Quest, A. L.: Supervised Study, 1916. Hennon, V. A. C. : The Relation Between Mode of Presentation and Retention. Psychological Review, 19: 79-96. Hillegas, M. B.: The Measurement of Quality in English Composition. Teachers' College Record, 1912. Huey, E. B. : Experiments in the Physiology and Psychology of Read- ing. American Journal of Psychology, 9: 575-586; 11: 283-302; 12: 292-312. Horn, Ernest: Principles of Method in Teaching Spelling as Derived from Scientific Investigations. Eighteenth Year Book, Part II; National Society for the Study of Education. Johnston, J. H. : Scientific Supervision of Teaching. School and So- ciety, 5: 181-188, 1917. Judd, C. H. : Practice Without Knowledge of Results. Psychological Review Monograph Supplement, No. 1, 1905. : Measuring the Work of the Public Schools, 1916. Kelly, F. J.: The Results of Three Types of Drill on the Fundamen- tals of Arithmetic. Journal of Educational Research, 2 : 693-700, 1920. King, I.: Comparison of Rapid and Slow Readers in Comprehension. School and Society, 4: 320-324, 1916. Kirby, T. J. : Practice in the Case of School Children. Teachers' Col- lege Contributions to Education, No. 58, 1913. Kirkpatrick, E. A.: An Experiment in Memorizing vs. Incidental Learning. Journal of Educational Psychology, 5: 405-413, 1914. Lakenan, M. E.: The Whole and Part Method of Memorizing Poetry and Prose. Journal of Educational Psychology, 4: 189-198, 1913. Lane, H. A. : Standard Tests as an Aid to Supervision. Elementary School Journal, 15: 378-386, 1915. Luqueer, F. L. : Self -Accounting in Supervision. Educational Re- view, 49: 460-468, 1915. 82 A Comparative Study of Mayman, J. E.: An Experimental Investigation of the Book, Lecture, and Experiment Method of Teaching Science in the Elementary Schools. Journal of Educational Psychology, 6: 246-250, 1915. McClelland, W. W. : An Experimental Study of the Different Methods of Subtraction. Journal of Experimental Pedagogy, 4: 293-299, 1918. McMurry, F. M. : How to Study, 1909. McGregor, Laura: Supervised Study in English, 1921. Mead, A. R. : Tendencies in Educational Measurements. Educational Review, 61: 117-127, 1921. Mead, C. D.: Spelling by Visualization Versus Drill Methods. Jour- nal of Educational Psychology, 5: 29-31, 1914. : Results in Silent Reading Versus Oral Reading. Journal of Educational Psychology, 8: 367-368, 1917. : An Experiment in the Fundamentals, 1921. Mead, C. D., and Johnston, C. W. : Testing Practice Material in the Fundamentals of Arithmetic. Journal of Educational Psychol- ogy, 9: 289-297, 1918. Minnich, J. H.: An Experiment in the Supervised Study of Mathe- matics. School Review, 21 : 670-675, 1913. Monroe, W. S., et al. : Educational Tests and Measurements, 1917. Monroe, W. S.: Measuring the Results of Teaching, 1918. Nutt, H. W. : Rhythm in Handwriting. Elementary School Journal, 19: 532-540, 1919. Nutt, W. N.: The Supervision of Instruction, 1920. O'Brien, J. A.: Silent Reading, 1921. Payne, G. E.: An Experiment in Motivation. Elementary School Journal, 17: 727-733, 1917. Pearson, H. C. : The Scientific Study of the Teaching of Spelling. Journal of Educational Psychology, 2: 241-252, 1911. Pechstein, A. L. : Alleged Element of Waste in Learning a Motor Problem. Journal of Educational Psychology, 8: 303-310, 1917. : Whole Versus Part Methods of Learning Nonsensi- cal Syllables. Journal of Educational Psychology, 9: 381-387, 1918. : Best Method of Mastering a Motor Problem. Ele- mentary School Journal, 17: 734-740, 1917. Peters, C. C. : The Influence of Speed Drills on the Rate and Effective- ness of Silent Reading. Journal of Educational Psychology, 8: 350-366, 1917. Peterson, J.: The Effect of Attitude on Immediate and Delayed Re- call. Journal of Educational Psychology, 7: 523-532, 1916. Phillips, F. M.: Value of Daily Drill in Arithmetic. Journal of Edu- cational Psychology, 4: 159-163, 1913. Pintner, R., and Gilleland, A. R. : Oral and Silent Reading. Journal of Educational Psychology, 7: 201-213, 1916. Pintner, R. : Oral and Silent Reading of Fourth Grade Pupils. Jour- nal of Educational Psychology, 4: 333-337, 1913. Pitman, M. S.: The Value of School Supervision, 1921. Proctor, W. M.: Supervised Study on the Pacific Coast. School and Society, 6: 326-328, 1917. Pyle, W. H.: Economical Learning. Journal of Educational Psychol- ogy, 4: 148-158, 1913. : Concentrated Versus Distributed Practice. Jour- nal of Educational Psychology, 5: 247-258, 1914. Pyle, W. H., and Snyder, J. C: The Most Economic Unit for Commit- ting to Memory. Journal of Educational Psychology, 2: 133- 142, 1911. Directed and Undirected Teaching 83 Reavis, W. C. : Factors That Determine the Habits of Study of Grade Pupils. Elementary School Teacher, 12: 71-81, 1912. : Importance of a Study Program for High School Pupils. School Review, 19: 398-405, 1911. Ross, C. : Problem of Supervised Study in the Grades. Education, 39: 457-470, 1919. Ruger, H. A.: The Psychology of Efficiency. Archives of Psychology, No. 15. Rugg, E. U. : Character and Value of Standard Tests in History. School Review, 27: 757-771, 1919. Rugg, H. O. : Statistical Methods Applied to Education, 1917. Saam, T. : Intelligence Testing as an Aid to Supervision. Elementary School Journal, 20: 26-32, 1920. Sachs, F.: A Preliminary Study of the Relative Efficiency of Induc- tive and Deductive Teaching of Logical Fallacies. Journal of Educational Psychology, 4: 381-391, 1913. Sandwick, R. L. : How to Study and What to Study, 1916. Schmidt, W. A. : An Experimental Study in the Psychology of Read- ing. Supplementary Educational Monograph, No. 2, 1917. Simpson, Mabel: Supervised Study in History, 1918. Starch, D. : Unconscious Imitation in Handwriting. Psychological Review, 18: 223-228, 1911. : Periods of Work in Learning. Journal of Educa- tional Psychology, 3: 209-213, 1912. : The Measurement of Efficiency in Reading. Jour- nal of Educational Psychology, 6: 1-24, 1915. : The Measurement of Efficiency in Writing. Jour- nal of Educational Psychology, 6: 106-114, 1915. : The Measurement of Efficiency in Spelling. Jour- nal of Educational Psychology, 6: 176-186, 1915. : The Measurement of Achievement in English Grammar. Journal of Educational Psychology, 6: 615-626, 1915. : Scale for Measuring Ability in Arithmetic. Jour- nal of Educational Psychology, 7: 213-222, 1916. : Educational Measurements, 1916. Stockton, J. L.: Effective Supervision. Education, 33: 344-351, 1913. Stone, C. W. : Problems in the Scientific Study of the Teaching of Arithmetic. Journal of Educational Psychology, 4: 1-16, 1913. : Standardized Tests in Arithmetic and How to Use Them. Teachers' College Contributions to Education, No. 83, 1916. Strayer, G. D. : Use of Tests and Scales of Measurement in the Ad- ministration of Schools. N. E. A. Report, 1915: 579-582. Suzzalo, H.: The Teaching of Spelling, 1910. Terman, L. M. : The Measurement of Intelligence, 1916. Thorndike, E. L. : An Improved Scale for Measuring Ability in Read- ing. Teachers' College Record. 16: 445-467, 1915, and 17: 40-67, 1916. : Mental and Social Measurements, 1913. : The Psychology of Drill in Arithmetic. Journal of Educational Psychology, 12: 183-194, 1921. : The Relation Between Speed and Accuracy in Addition. Journal of Educational Psychology, 5: 537-544, 1914. : Practice in the Case of Addition. American Jour- nal of Psychology, 21 : 483-486, 1910. : The Curve of Work. Psychological Review, 19: 165-194, 1912. Trabue, M. R.: Completion Test Language Scales. Teachers' College Contributions to Education, No. 77, 1916. 84 A Comparative Study of Uhl, W. L. : The Use of the Results of Reading Tests as Bases for Planning Remedial Work. Elementary School Journal, 17: 266- 275, 1916. : The Use of Standardized Materials in Arithmetic for Diagnosing Pupils' Methods of Work. Elementary School Journal, 18: 215-218, 1918. Vance, W. M.: How Shall the Superintendent Measure His Own Effi- ciency? N. E. A. Report, 1914: 279-283. Wagner, C. A.: Supervision of Instruction. Educational Review, 59: 137-141, 1920. Whipple, G. M.: How to Study Effectively, 1916. Wilson, H. B.: Training Pupils to Study, 1918. Wilson, G. M., and Hoke, K. J.: How to Measure, 1921. Willing, M. A. : Measurement of Written Composition in Grades VI to VIII. English Journal, 7: 193-202, 1918. Winch, W. H. : Additional Researches in Learning to Spell. Journal of Educational Psychology, 7: 39-110, 1916, : Inductive Versus Deductive Methods, 1913. Woody, C. : Measurement of Some Achievements in Arithmetic. Teach- ers' College Contributions to Education, No. 80, 1916. Zimmers, P. J.: Teaching Boys and Girls How to Study, 1918. Zirbes, Laura: An Experimental Evaluation of Method in Spelling. Elementary School Journal, 19: 778-798, 1919. Diagnostic Measurement as a Basis of Procedure. Elementary School Journal, 18: 505-522, 1918. LIBRARY OF CONGRESS 019 745 424 6