a^ 5( • ervi^ion in. of rtje Wim\}tt^\tV of i?ortf) Carolina Wi)ii boofe toag presientetr Educational Publication No. 106 Division of Supervision No. 25 A Study of the Value of Supervision in Consolidated Schools BY MAYCIE SOUTHALL Assistant Supervisor of Rural Elementary Schools published by the State Superintendent of Public Instruction Raleigh, N. C. RALEIGH Edwards & Broughton Company 1925 CONTENTS Page Letteb of Transmittal 5 Preface ; ._ 6 Abstract 7 Tables 8 Diagrams .— 9 Chapter I. Introduction .- - 11 Chapter II. Factors for Wliicli Control was Attempted 13 Chapter III. Effect of Supervision upon Factors Measured 21 1. Improvement in Reading 22 2. Improvement in Spelling 24 3. Improvement in Four Fundamentals 26 4. Improvement in Problem Solving 28 Chapti-:r IV. Summary of Results and Conclusions 31 1. Comparison of Progress by Subjects ...^ 31 2. Comparison of Progress by Grades -32 3. Summary of Results .,._. 34 4. Conclusion 34 Appendi-X Statistical Data 35 LETTER OF TRANSMITTAL October 1, 1926. State Superintendent A. T. Alxen, Raleigh, N. C. Deab Mr. Allen: Is more efficient and intensive supervision of our rural schools a present urgent need? Does expert supervision enable pupils to do more and better work? Is it really needed in our long-term con- solidated rural schools? Is it worth the cost? The foregoing are some of the questions now being asked more frequently than before by progressive county superintendents, county boards of educa- tion, county boards of commissioners, as well as by the individual taxpayer. And it is self evident that these vital questions demand an intelligent answer, one not the expression of horseback opinion, but one that is based upon indisputable facts acquired at first hand. Several successful attempts have been made to answer these questions. Just a few years ago, in Brown County, South Dakota, M. S. Pittman effectively demonstrated the value of intensive rural school supervision in the one and two-teacher types of school. In 1923-25, under the direction of State Superintendent Burris, the eifectiveness of intensive supervision in promoting the learning of the pupils was clearly demonstrated using all types of schools in LaGrange and Johnson counties of the State of Indiana. Recently a similar experiment has been worked out in the rural schools of Oakland and Macomb counties, Michigan, under the direction of Dr. Ernest Burnham. Here again it was clearly demonstrated that intensive and expert supervision of rural schools pays. But notv/ithstanding the fact that the value of expert supervision of rural schools in promoting the learning of the pupils, in small schools had been demonstrated beyond the shadow of a reasonable doubt in other states, yet it seemed advisable to make a thorough-going, first-hand study of this sub- ject here in North Carolina by using larger schools. Therefore, in 1924-25, Miss Maycie Southall, Assistant State Supervisor of Rural Schools, undertook to ascertain the value of expert supervision of instruction in our larger type of consolidated rural schools, with an eight- months school term, using for her demonstration Craven and Jones counties. And so thoroughly did Miss Southall work out this experiment that the George Peabody College for Teachers gladly accepted this piece of work for her Master's Thesis. Believing that this bulletin will prove of definite value to county superin- tendents, rural school supervisors, rural school principals, county boards of education and to Departments of Education in our various institutions of learning, I feel justified in asking that you have it printed in bulletin form for distribution. Yours truly. a^' ^ I >^(?-^d^2-'^-^ State Supervisor of Rural Schools. PREFACE This study is reported with the hope that the findings may prove of service in promoting more adequate and effective supervision of instruction, especially in the rural schools of North Carolina. It was made upon the assumption that supervision is valuable to the extent that it aids pupils to do better school work than they would otherwise do. Only such phases of pupil progress as could be measured objectively are reported herein. Grateful acknowledgment is made to the county superintendents, the prin- cipals and the teachers of both the supervised and the control groups of schools whose hearty cooperation made the investigation possible. Special recognition is made of tlie services of Miss Margaret Hayes, the supervisor of the supervised group of schools, and the corps of teachers who worked so zealously with her; also, Mr. Carl S. Adams, Professor of Education, East Carolina Teachers College, and the group of students who scored the test results. A review of the statistical treatment of the results and helpful suggestions were given by Dr. Norman Frost, Professor of Rural Education, George Peabody College for Teachers. To all who contributed in any way to this investigation the members of the Division of Supervision express their sincere appreciation. Maycie Southall, Assistant Bupervisor of Rural Elemejitary Schools. October 1, 1926. ABSTRACT The Problem. Is there a need for, if so what, is the value of supervision In the rural consolidated schools of North Carolina? Since it was necessary to limit the scope of the investigation, it was re- stricted to a study of supervision in the consolidated schools for two reasons: first, because of the present state-wide consolidation program and second, because the need for supervision in schools of this type is not so generally recognized as it is for those of the one, two and three-teacher type. Tlie Method Used. The equivalent group method was considered best suited to an investigation of this kind. One group of the schools selected for this study was located in a county with initial supervision while the other was in a county having no special supervisory officer. In summing up the equivalence of the two groups at the beginning of the experiment, Chapter II, it was found that with the exception of supervision, there were very slight differences in any of the important factors, and that such differences as there were tended to counteract each other in their partiality, with a slight favoritism to the control group. The Eesults and Conclusions. The extent and particulars in which supervision gave positive results are enumerated in Chapter IV. Briefiy stated, the initial tests revealed that the previous achievement of both groups of pupils was below normal in every subject. This would indicate a need for supervision of instruction or some agency for increasing teaching efficiency. During the five months studied, the children in the group of supervised schools made more than normal progress in every subject measured, with an average progress of 2.26 times as great as that of the control group. On a basis of the time spent in these schools, one supervisor could supervise thirty-six teachers. On a basis of the results obtained, and the current cost of instruction, the services of a supervisor who could produce such improve- ment in the total results of thirty-six classrooms would be valued at $12,474. From the findings of this limited investigation, it seems fair to conclude that there is a need for supervision in the consolidated schools, and that even if the tool subjects alone were affected thereby, such supervisory services would be worth their cost. TABLES Page Table I. A Comparison of the Educational Status of ttie Two Groups as Shown by the Fall Test Scores .- — 14 Tahle II. A Comparison of the Mental Status of the Two Groups as Shown by the National Intelligence Test, Scale A 18 Tahle III. Reading, A Comparison of the Progress of the Two Groups Over a Five Months' Period 22 Table IV. Spelling, A Comparison of the Progress of the Two Groups Over a Five Months' Period 24 Table V. Four Fundamentals, A Comparison of the Progress of the Two Groups over a Five Months' Period 27 Table VI. Reasoning, A Comparison of the Progress of the Two Groups over a Five Months' Period 29 Table VII. Comparison of the Progress of the Two Groups by Subjects, Based upon the Combined Scores in Grades IV-VII 31 Table, VIII. A Comparison of the Progress of the Two Groups by Grades Based upon the Combined Scores in Readiirg, Spelling, Computation and Reasoning 32 DIAGRAMS Page Diagram 1. Comparing the Achievement of the Two Gi'oups in Reading and Spelling with the Standard Norms at the Beginning of the Experiment 16 Diagram 2, Comparing the Achievement of the Two Groups in Arith- metic with the Standard Norms at the Beginning of the Experiment — 17 Diagram 3. Comparing the Native Ability of the Two Groups at the Beginning of the Experiment 19 Diagram 4. Illustrating the Improvement in Reading over a Five Months' Period, as Shown by the Thorndike McCall Read- ing Scale 23 Diagram 5. Illustrating the Improvement in Spelling over a Five Months' Period as Shown by the Morrison McCall Spelling Scale .... 25 Diagram 6. Illustrating the Improvement in the Fundamentals of Arith- metic over a Five Months' Period as Shown by the Woody ' Arithmetic Scale, Series B 28 Diagram 7. Illustrating the Improvement in Reasoning in Arithmetic over a Five Months' Period as Shown by the Buckingham Reasoning Scale 30 Diagram S. Comparison of the Improvement of the Two Groups over a Period of Five Months, Expressed in the Decimal Part of a Normal Year's Progress 32 Di.\GRAM 9. Illustrating the Per cent of Normal Progress Made by Each Grade, Based upon the combined Reading, Spelling and Arithmetic Scores 33 A STUDY OF THE VALUE OF SUPERVISION IN CONSOLIDATED SCHOOLS CHAPTER I OTRODUCTIOF Purpose of the Investigation The purpose of this study was to determine objectively the effect of county supervision upon certain factors of school efficiency. The data thus ob- tained is used to measure tlie practicality of supervision and furnish a basis for determining the advisability of joint county and state support of rural supervision in North Carolina. A brief explanation is made to show why the study was undertaken. Sec- tion 51, of the last codification^ of the public school law of North Carolina, permits a County Board of Education to employ a supervisor or supervisors to aid the county superintendent in supervising instruction: provided, the salary of the same is stipulated in the budget and approved by the County Commissioners. A part of the salary of the supervisor may be paid out of the State Public School Fund: provided, the duties of the same are approved by the State Superintendent of Public Instruction. In practice, the State has usually paid one-half of the rural supervisor's salary, exclusive of traveling expcrses. Even with this much help from the State the counties have been slow in appropriating their part. At present, twenty-six, or approximately one-fourth of the counties, have a rural supervisor. In several instances the County Board of Education has placed their proportional part of the super- visor's salary in the May budget, but the appropriation failed to be approved by the County Commissioners. This is not surprising when we consider that the County Commissioners are usually business men wholly out of touch with the work of the schools. In a few of the mountain and coast counties they have hardly become accustomed to the need for a whole-time superintendent. Therefore, it is quite natural that they should question the advisability of this extra expenditure for the supervision of instruction unless the county superintendent and his Board can give very convincing evidence that supervision is essential to effective teaching results. M. S. Pittman^ showed in his experiment in the one and two-teacher schools of Brown County, South Dakota, that the "Zone Plan" of supervision was a very postive factor in the improvement of instruction in small isolated schools. The State Department of Education in Indiana'- with the financial aid of the General Education Board, has demonstrated the value of county- wide supervision in LaGrange and Johnston counties, using two special supervisory officers in each county, over a two-year period. Although the results of both these investigations were highly favorable for supervision, the Director of Rural Supervision in North Carolina thought that local ^ The Public School Law of North Carolina, Codification of 192:^, State Department of Public Instruction, Raleigh, N. C, p. 17. ^ Pittman, M. S. — The Value of School Supervision, Demonstration with the Zone Plan in Rural Schools, Baltimore, Warwick & York, 1921, pp. 94-108. ^ "Preliminary Report on Supervision in County Demonstration" Indianapolis, Indiana. Department of Public Instruction, Bulletin No. 74, 1924, p. 37. 12 ■ Value of Supervision evidence, if such could be secured, would help very materially in convincing the authorities whose financial support it is necessary to secure in order to introduce rural supervision in the other seventy-four counties. Therefore, this study was undertaken to obtain an objective valuation of rural school supervision as it is administered in North Carolina, and to find whether or not it is worth its cost. Since the consolidated school is fast becoming the permanent type of rural school in North Carolina and since one teacher per grade is the minimum for a standard elementary school,* and since there is already very convincing evidence of the value of supervision in the small ungraded schools,' it was considered best to restrict the study to the following problem: Is there a need for, and if so what is the value of supervision in the rural consolidated and small town schools? MetLod of Study Used The ejiuivalent group method was selected for conducting the experiment. Therefore it was necessary to find two groups of consolidated schools practically equivalent in all particulars except supervision. There were numerous counties from which a group of unsupervised schools might be selected, but only one, Craven County, in which the effect of an initial supervisory program could be investigated. The Craven County superintend- ent, Mr. R. S. Proctor, expressed a willingness to cooperate in the experiment; and the supervisor. Miss Margaret Hayes, agreed to follow the supervisory program worked out by the Division of Supervision and have the effect of her first year's work measured. The Dover, Vanceboro, and Ernul Schools of this county were selected as the Supervised Group. The next step was to find another group of schools very much like the supervised group, so that there would be no advantages to either except in the one factor, supervision. Such a group was found in the Trenton, Pollocksville, and Maysville Schools of Jones County. With the permission of the county superintendent and the principals of these schools, they were made the Control Group. Upon the assumption that reading, arithmetic, and spelling have a social value; and upon the further assumption that the elementai-y school should give its pupils a mastery of the tool subjects, standard achievement tests were given in these subjects. The educational results, as shown by such tests, are used as reliable data from which to draw conclusions as to the value of supervision. The progress of the two groups of children was measured over the same five months' period in the following functions: 1. Ability to comprehend what is read in a long time diflSculty test. (Thorndike-McCall Reading Scale, Forms I and III.) 2. Ability to spell commonly used words. (Morrison-McCall Spelling Scale. List;) I and VIII.) 3. Ability to add, subtract, multiply and divide integers, common and decimal fractions, and denominate numbers. (Woody Arithmetic Scale, Series B, Forms I and II.) 4. Ability to reason out solutions of one, two, three and four-step problems. (Buckingham Reasoning Scale, Forms I and II.) ^ Publications of the State Department of Public Instruction, Raleigh, N. C. "Consolida- tion of Schools," State School Facts, Vol. 1, No. 21, January, 1925; "Public School Law," Goditication 1923, Article 6, Sec. 73, pp. 23-28; "Elementary School St.andards in North Carolina," 1924. ^ Idem.. Pittman. M. S., p. 107. CHAPTER II FACTORS FOR WHICH CONTROL WAS ATTEMPTED Since the equivalent-group method was to be used in making this study, it was necessary that every precaution be taken to foresee the constant irrele- vant factors and to eliminate or equate the most significant ones. This was done by selecting two groups of schools that were equivalent or prac- tically so. :#^^ In order to determine the equivalence of the schools which formed the groups, comparisons were made in the following particulars: 1. Type of schools. 2. Length of school term. - 3. Teaching skill — training, experience, and tenure of the teachers. 4. Teaching load — enrollment and number of grades. 5. Educational status of the children. 6. Mental status. 7. Other factors: Nationality of children, supervision by county superin- tendent, administering the tests, and scoring the results. Types of Scho-ols. In the original plan, the supervised group contained three schools with four, seven, and eight elementary teachers respectively, while the control group had three schools with four, seven and nine elemen- tary teachers. Near mid-term, an extra teacher was placed in the four-teacher school of the supervised group. Since this addition gave the supervised group a decided advantage, the four-teacher schools were dropped from the study. They were not retested nor considered in any part of this report. This left the experimental grades, four to seven, with one, and only one teacher per grade in each of the schools of both groups. Therefore, in this particular there was no advantage to either group. The school buildings of .both groups were of brick, and were modernly constructed. With the exception of two rooms in the supervised group and one in the control group, thejf were equipped with single patented desks. The instructional equipment Avas -very much alike both in kind and amount. Length of Scliool Term. Since all of the schools had an eight months' school term, there was no advantage to either gToup. The schools of the control group opened later in the fall, but this was taken care of in planning the testing program. It was so planned that the first series of tests were given during the last week of the second month and the second series during the last week of the seventh month. Thus, the same five months' period was measured in each school of both groups. Teaching Skill. The relative ability or teaching skill of the two groups was equated on the basis of training and experience. These were taken as a fair measure because the certification of teachers in North Carolina is based upon training, and the minimum salary is based upon the certificate held, with allowance for each year's experience through four years. Tenure was also taken into account in measuring teaching ability. According to the certificates held, the training of the supervised group averaged four years of high school and 1.12 years of collegiate training; while the control group averaged 1.25 years of collegiate training above high school. A slight ad- vantage of .13 of a collegiate year was in favor of the control group. The control group also had the advantage in teaching experience. The average experience of the supervised gi'oup was 3.4 years, including one in- 14 Value of Supervision experienced teacher; while tlie average experience of the control group was over four years and did not include any inexperienced teachers. In tenure, the control group again had the advantage. Of this group, 87.5 per cent had taught in the same school the previous year while only 62.5 per cent of the supervised group were teaching in the same school for the second year. Thus we see that in all three, training, experience, and tenure of its teachers, the control group had a slight advantage. Teadiing" Load. The grades selected for study had one and only one teacher per grade in each school. The average enrollment per teacher was 30.1 in the supervised group, while that of the control group was 33.9. This gave the teachers of the supervised group an advantage of 3.S less pupils in average enrollment. Of the average number enrolled, an average of 1.79 pupils in the control group, and 1.46 pupils in the supervised group were children of borderline intelligence. Although these children were excluded before measuring the progress of the grades, still it gave the teachers of the supervised group a slight advantage in the daily instruction. Thus we see, that in teaching load, whatever advantages there were favored the supervised group. Educational Status of (croups. The fall test scores were used as a com- parative measure of the ability of the two groups to make progress as based upon former scholastic attainmeni. Since the acquired ability or educational status of children is one of the two most important factors in determining their ability to make progress, it was most fortunate to find the two groups practically equivalent in all subjects and grades tested. In order to make a direct comparison of the achievement in the different subjects, the various test scores were translated into a common unit of measure. The grade score was selected for this purpose because the school grade is probably the most widely recognized term for expressing pupil prog- ress. A brief explanation of the method used to convert test scores into grade scores follows. The grade scores were obtained by writing the author's norms at the approximate place on the McCall Multi Mental Scale and then interpolating and extending the tables graphically. A copy of the table thus computed and used in this experiment to change T scores or crude scores into grade scores, is listed as Table IX in the Appendix. It will be noticed that, for each test in the various subjects, standards were determined for each grade and tenth of a grade. By referring; the actual test score to this scale of grade standards, the grade score is found at the grade and tenth of a grade for which this score is the average achievement of pupils. TABLE I A COMPARISON OF THE EDUCATIONAL STATUS OF THE GROUPS AS SHOWN BY THE FALL TEST SCORES Thorndike McCall Reading Scale TWO T Scores: Supervised Group Control Group *Grade Scores: Supervised Group. Control Group 'Difference in Grade Units IV V VI VII 33.6 40.6 43.0 49.2 35.0 40.5 45.2 48.0 3.5 4.6 5.1 6.1 3.7 4.6 5.4 6.0 .2c .0 .3c .1 IV-VII Mean .Ic In Consolidated Schools 15 MoEKisoN McCall Speilling Scale Crude Scores: Supervised Group. Control Group Grade Scores: Supervised Group, Control Group Difference in Grade Units- 17.4 21.3 28.0 31.0 19.9 21.9 28.9 35.6 3.4 4.1 5.2 5.8 3.9 4.2 5.4 6.8 .5c .Ic .2c 1.0c The "Woody Aeithmetic Scale, Series B Crude Scores: Super\ased Group. Control Group Grade Scores: Supervised Group. Control Group Difference in Grade Units. 24.7 35.5 42.2 48.2 28.8 34. 1' 43.4 46.0 3.4 4.2 5.1 5.9 3.6 4.1 5.3 5.7 .2c .la .2c .2s The Buckingham Re.\soning Scale Crude Scores: Supervised Group Control Group Grade Scores: Supervised Group Control Group Difference in Grade Units. Mean Grade Score: Supervised Group Control Group Difference in Grade Unite. 36.2 49.2 53.6 70.6 36.3 49.0 53.2 67.8 3.3 4.6 5.1 6.8 3.3 4.6 5.0 6.5 .0 .0 .Is .38 3.4 4.4 5.1 6.2 3.6 4.4 5.3 6.3 .2c .0 .2c .le .Ic Note. 'A grade score of 3.5 should be interpreted to mean the fifth month of the third grade. The letter c is used when the difference favors the control group and the letter s when it favors the supervised group. Table I gives both the crude and grade scores made in the initial test. The difference in the status of the two groups at the beginning of the experiment, is expressed in grade units or the decimal part of a school grade. By reading the mean differences, we see that the control group had .1 of a grade better preparation in reading, and .5 of a grade better preparation in spelling. In the four fundamentals, the difference was zero, while in reason- ing in arithmetic the supervised group showed .1 of a grade better prepara- tion. "When the preparation of two groups is considered by grades, the control group had an advantage of from .1 to .2 of a grade in all except the fifth grade where the difference was zero. "When all subjects and grades were combined, there was an average difference of .1 of a grade in favor of the control group. In other words, in so far as attainment in reading, spelling and arithmetic was measured, the children of the control group had a greater mastery of these tool subjects and thereby a slight advantage in their preparation for these and related subjects. As both groups of children had already been in school two months when the fall tests were given, the grade norm for the fourth grade would be approximately 4.2, the norm for fifth grade 5.2, etc. Diagrams 1 and 2 show that neither group of children had made normal achievement in any subject or grade. By comparing the mean grade scores at the bottom of Table I with the norms for the respective grades, we find that the fourth and fifth 16 Value of Supebvision grades of the supervised group were .8 of a grade below standard, the sixth 1.1 and the seventh grade 1.0 grades below standard; in the control group, the fourth grade was .6, the fifth grade .8, and the sixth and seventh grades each .9 of a grade below standard in achievement. When all grades were Or/\D£ •Scofl£S 7.0 6.0 S.O 4.0 3.0 2.0 1.0 £/=? I t^E^/VT S/'y/rn THE STAA/DAffD considered together, the supervised group was found to be on an average of .9 of a grade below standard, and the control group .8. This would indicate that for the last few years at least, neither group had made normal progress, and that they were both nearly a school year below standard in the subjects tested. . ■ • . In Consolidat-ed Schools 17 I The above facts lead us to conclude that in so far as acquired ability wag concerned, the two groups were practically equivalent. If either had an advantage in previous attainment or preparation for the work of the respect- ive grade, it was the control group. Gf^Aoe 6 s.o 4 3 O 3 Grade Sco/?£s 7 e o 9:0 3.0 ^ O /.O rC/WAM^A/TAlS C/?AD£lZ Orade'Y: Graob'W. /^^ASOA/JA/6 LI /VO/RM. SUPR. Gf?0(yP CO/VTROL 6/=fOO/=> CVI >f vO Grace 7K GaADe:3: FfADS iss: o/^AofWr K COM3 O RAD£ \JI/ 0/A "y ^ k. qJ 1^ Q ^ ^ o fifty words each. The lists were so chosen from the Ayres Scale of 1,000 words as to be of uniform difficulty, and composed of words progressively ranging from easy to difficult. The First List of this scale was used for the fall test, and the Eighth List for the spring test. To prevent any com- 26 Value of Supervision munication of words, all grades were tested in spelling between the same intermission periods and in the order of their rank. To prevent any mis- understanding or confusion of words, each was pronounced, used in a sentence, and pronounced the second time. For example, "company, We are expecting company for dinner, company." The average number of words spelled correctly is given by grades for the test given at the end of the second month, and for the one five mouths later. As in reading, the spelling improvement is expressed in grade units or the decimal part of a school grade completed in five months. The sum- mary is given in terms of the per cent by which the progress of one group exceeds the progress of the other. It will be observed from the above table and the accompanying graph that in the fall test, the typical spelling performance of the pupils in the control group was better in every grade than the spelling in the supervised group, but that in every grade the average improvement was greater in the supervised group. When compared with the progress to be expected the supervised group made normal progress or above in every grade except the fourth while the control group made normal progress only in the seventh grade. The average progress of the supervised group exceeded the average progress of the control group by 182 per cent. In other terms, the children where there was supervision of instruction, made 2.82 times as much progress in spelling as did those of the control group. On the basis of the above statements, it is concluded that the supervision of the instruction very materiallj' affected the efficiency of the teaching of spelling. Section 3 — Improvement in tlie Four Fundamentals The purpose of this section is to answer the question: What was the effect of supervision upon the efficiency of the teaching of the fundamental operations in arithmetic? The test used to measure tlie extent to which the children of the two groups had become automatic in the fundamental processes was the Woody Arith- metic Scale, Series B. This is a four-page booklet testing their ability to add, subtract, multiply and divide integers, common and decimal fractions, and denominate numbers. The problems on each page range in difficulty from those so simple that any third grade can solve them, up to others too difficult for the average eighth grade to solve. It presents nineteen different types of difficulty in addition, fifteen in subtraction, twenty in multiplication, and fifteen in division. Thus by this series of problems, the pupil is tested more or less over the entire range of fundamental processes. Since accurate computation in arithmetic in the upper grades demands a mastery of all four of the fundamentals, they will be considered together rather than separately in this comparison. Table V gives the average number In Consolidated Schools 27 of problems worked correctly by each gi-ade, out of a possible sixty-nine, in the fall and in the spring. It also gives the grade scores with the im- provement expressed in grade units. TABLE V .FOUR FUNDAMENTALS, A COMPARISON OF THE PROGRESS OF THE TWO GROUPS OVER A FIVE MONTHS' PERIOD Groups and Grades Crude Scores Grade Scores Imp. in Fall Spring Fall Spring GradeUnits Supervised Group: Grade IV 24.71 35.49 42.19 48.21 33.75 41.49 52.54 55.41 3.4 4.2 5.1 5.9 4.1 5.0 6.4 7.0 .7 .8 1.3 Grade VII 1.1 98 Control Group: Grade IV 28.83 34.10 43.36 46.00 34.39 35.18 46.97 49.00 3.6 4.1 5.3 5.7 4.1 4.2 5.8 6.0 Grade V . . ._ .1 Grade VI .5 Grade VII - .- - .3 1 .37 Progress of the supervised group exceeded the progress of the control group 165 per cent. (Appen- dix, Table III.) Table V and the accompanying graph show that at the first testing, the control group had better averages in the fourth and sixth grades while the supervised group had the better scores in the fifth and seventh grades. In the test given five months later, the scores for the supervised group were greater in all grades except the fourth. In every case the supervised group made greater progress in the fundamental operations than the control group. The supervised group made an average total improvement pf .98 of a grade as against .37 for the control group. If the total progress of the supervised group is expressed in terms of the progress of the control group, it is 265 per cent or, in other terms, the children where the instruction was supervised gained 2.6-" times as much skill in rapid and accurate com- putations as did those in the unsupervised group. From the above data it seems fair to conclude that supervision was a positive factor in influencing the efficiency of the teaching of the fundamental operations in arithmetic. 28; Value of Supervision ts M f° ■ ><0 <\i UQ ^ ^ o Section 4 — Improvement in Reasoning or Problem Solving The purpose of this section is to answer the question, What was the effect of supervision upon the teaching of reasoning, or problem solving in arithmetic? Helping pupils acquire the ability to reason independently is usually con- sidered an important and rather difficult part of the teaching of arithmetic. In order to show some direct comparison of the improvement in the problem- solving ability of the two groups, the Buckingham Scale for Problems in Arithmetic was given. Although it is broken up into three divisions, they are so arranged that all pupils of grades three through eight may be measured on the same scale. For example, the last page of problems in Division I is the same as the first page in Division II, and the last page in In Co?rsoLii>ATED Schools 29 Division II is the same as the first page in Division III. Although this overlapping allows for considerable range of ability within the grades, the range within the grades tested was greater than the tests allowed, causing several to fail to score on the fall test. These eliminations left the number tested in reasoning smaller than in any other subject. TABLE VI REASONING, A COMPARISON OF THE PROGRESS OF THE TWO GROUPS OVER A FIVE MONTHS' PERIOD Groups and Grades Crude Scores Grade Scores Imp. in FaU Spring Fall Spring GradeUnits Supervised Group: Grade IV 36.24 49.22 53.63 70.63 43.86 60.42 62.11 78.36 3.3 4.6 5.1 6.8 4.2 5.7 5.9 7.6 .9 Grade V 1.1 Grade VI .8 Grade VII .8 .90 Control Group: Grade IV 36.26 49.00 53.19 67.75 42.46 53.17 61.19 72.42 3.3 4.6 5.0 6.5 4.0 5.0 5.8 6.9 7 Grade V .. .4 Grade VI . . . .8 Grade VII , .4 Total Av. Improvement.. .58 Progress of the supervised group exceeded progress of the control group 55 per cent. (Appendix Table IV.) Table VI gives by grades the average fall scores, spring scores and average improvement in terms of grade units for each group. From this table and Diagram 7 it is obvious that the supervised group made more progress than the control group in every grade except the sixth, where the difference was zero. It should not be overlooked, however, that the fourth and sixth grades of the control group made more than normal progress, and that as a whole, the control group made greater progress in problem solving than any other subject measured. Nevertheless, the children of the supervised group made 55 per cent more improvement in problem solving than the control group. It seems from the facts presented in the summary of results by subjects, Table VII, that the teachers of the supervised group in general achieved better results in teaching the fundamental operations in arithmetic than they did in problem solving. The latter is more difficult to teach, more difficult to find satisfactory drill exercises for, and probably even more de- pendent upon the native capacity of pupils. 30 Value of Supebvisioiv Since in all grades except the seventh there was less than one unit of difference in the fall test, and since in every grade there was greater im- provement in the supervised than in the control group, and since the super- vised group made an average improvement of one and one-half times as much as the control group, it is concluded that the supervision of instruction did affect the efficiency of the teaching of reasoning or problem solving. I N H N \- IN « Q 'I c^ ^ CHAPTER IV SUMMARY OF RESULTS AND CONCLUSIONS Comparison of Progress by Subjects The purpose of this section is to show which group of schools made the greater progress in each suhject and in which subjects the greatest progress was made during the five months' period of the experiment. Progress is expressed in the decimal part of a normal year's work with tables and graphs used to illustrate the comparison. Since all of the schools in both groups had an eight months' term, each group would be expected to do five-eights or .62 1/^ of the normal work of a school grade. In Tables I- IV of the Appendix will be found the average or mean results of each subject by grades. Table VII of this chapter gives the combined results for all grades, which is the mean of the separate grade averages. TABLE VII A COMPARISON OF THE PROGRESS OP THE TWO GROUPS BY SUB- JECTS, BASED UPON THE COMBINED SCORES OP GRADES IV-VII Subjects Der-imal Part of the Normal Work of a School Grade Completed in Five Months or .62)2 of a School Year. Per cent by which the Prog- ress of the .Supervised group Exceeded the Progress of the Supervised Control Control. .63 .93 .98 .90 .25 .33 .37 .58 152% Spelli ng 182% Arith.— Fund Arith. — Reas 165% 55% Average . .86 .38 126% The data obtained from this study shows that: 1. When the results were considered for all the grades, the supervised group made more than normal progress in every subject measured, while the con- trol group made less than normal progress in all subjects. (A comparison of the fall scores with the standard grade norm showed that no grade of either group was standard in any subject. This would indicate that neither group had made normal progress in the last few years at least.) 2. The progress of the supervised group more than doubled the progress of the control group in reading, spelling, and the four fundamentals. 3. Both groups made less progress in reading than in any other subject. In the supervised group the greatest progress was in fundamental processes in arithmetic, while in the control group the greatest progress was in reason- ing or problem solving in arithmetic. 4. Out of sixteen differences before the results for all grades were combined, fifteen favored the supervised group. 5. The supervised group made normal progress or above in all subjects except sixth grade reading and fourth grade spelling. The control group made less than normal progress in all subjects except fourth grade reading, seventh grade spelling, fourth and sixth grade reasoning. 6. In the sixteen comparisons of achievement, the supervised group made normal progress or above fourteen times; the control group only four times. 32 Value of Supervision R£AD/h/6 SPELUNC APITHMETIC .98 FUNDAMENT/^LS :5T .62 s. A Rl THME TJC 3 O RFASON/NC .^O K////// NOffMAL PROGR£SS SUPR- GROUP CONTROL GROUP D//\6ffA/^ 8- Coy^PAfi>/SO/V OF IMPROVE: /^ENT iN TH£ rwo CffOUfS- OVER A rivs MO/VTRS' Pe^fOO £XfP?CS3£D IN TH^ DECIMAL /=>ART OF- A rvO/^A/iAL YEARS PROGRESS. Comparison of Progress by Grades The purpose of this section is to show which group of schools made the greater progress in each grade and to show in which grades the greatest progress was made. The comparisons are made in terms of the decimal part of a normal year's work, the per cent of normal progress made and the per cent by which the progress of one group exceeded the progress of the other. TABLE VIII A COMPARISON OF THE PROGRESS OF THE TWO GROUPS BY GRADES, BASED UPON THE COMBINED SCORES IN READING, SPELLING, COMPUTATION AND REASONING Grades Decimal Part of the Normal Work of a School Grade Completed in Five Months or .6232 of the School Term. Per cent by which the Prog- ress of the .Supervised group Exceeded the Progre.ss of the Supervised Control Control IV V VI VII .73 .88 .83 1.00 .53 .18 .43 .38 38% 389% 93% 163% Average .86 .38 126% In CONSOLIDAl-ED SCHOOLS 33 A comparison of the progress made by grades shows that: 1. In 62^ per cent of the school term, the supervised group accomplis.hed 86 per cent of a normal year's work; the control group 38 per cent. 2. No grade of the control group made normal progress when all subjects were combined, while no grade of the supervised group made less than normal progress. 3. Of the fourteen comparisons in which the supervised group made normal progress, four were in seventh grade, four in the fifth, three in the fourth, P£P GfiADE CENT HZ 8'5 S- Z9 •sar 69 szr Gf m-\m 61 'NOR, PROGRESS JOO PER GRAD£ CENT ET //7 JZ 141 YL t35 m /60 JF-EZT J38 CONTROL GROUP SUPERVISED GROUP NOR. PROGRESS /OO Z^ SO 7j5' /OO /Q.S /SO //^ 0/Aefi>flM9' /LLusTRArtf>/c THB pea C£NT or NOffMAL PROG/=>ESS UAD£ BY EPiCH GRADlz, BASED UPON THE COMBINED PPeAOlNG. SP^LU/^C, AND ARITHMETIC SEPIBS and three i-n the sixth. Of the four times that the control group made normal progress or above, two were in the fourth grade, one in the sixth, and one in the seventh. (See Tables I-IV of the Appendix.) 4. In the supervised group, the greatest progress was made by the seventh grade with rather uniform progress in the other three; in the control group, the greatest progress was made by the fourth grade and the least progress by the fifth. 5. In every grade tested, the supervised group made the greater progress. The greatest difference, 389 per cent, was in the fifth grades and the least difference. 38 per cent, was in the fourth grades. 34 Value of Supervision Slim mar J of Results This study shows that: 1. On an average the children of the supervised group in the five months period, advanced 126 per cent faster than the cliiklren in the control group, or 2.26 times as fast. 2. In 100 days, the children of the supervised group on the basis of the subjects measured received the average equivalent of 138 days of instruc- tion; the control group the equivalent of 61 days. 3. On a basis of the same difference for the entire term of 160 days, the children of the supervised group would accomplish the equivalent of 1.4 , grades (221 days) while the control group would accomplish .6 of a grade (98 days). 4. Within the time at the disposal of all rural children in North Carolina, 7-14 years, adequate supervision of instruction would help the children obtain an education 2.26 times as good as the one now being received. 5. If the compulsory attendance law required the completion of standard elementary grade work, at the rate of progress of the supervised group it could be completed in 5.7 years, or a saving of 2.3 years for work or higher education and a proportional saving in the cost of instruction to the taxpayer. 6. Put in other terms, one county for an additional expenditure of $350 for thirty-five days of supervision, purchased the equivalent of seventy-seven days of instruction. At the current daily cost of instruction in the control group this would have a monetary value of $2,772. 7. Upon the same time allotment^ one supervisor could supervise thirty- six teachers'. The services of one supervisor who could produce such results in the total results of thirty-six classrooms would be valued at $12,474. Conclusion It is generally conceded that mastery of the fundamental tools of learn- ing is one of the objectives of the elementary school. The results of this study indicate that the supervision of instruction was certainly a positive factor in accomplishing this objective. Further investigation is needed to determine the degree to which this is true of other important but less tan- gible aims of elementary education. However, within the limits of this ex- periment, it seems fair to conclude that supervision is a valuable means of improving teaching efficiency in the consolidated schools; and that even if it only affected the results in the tool subjects, it would seem to justify its present cost to the State and county school systems of North Carolina. ^ This is an understatement both of the amount of instruction purchased and the super- visoi'y load, because thirty-five days was the total time spent in supervising all the ele- mentary grades, yet it has been charged against only grades four through seven. This was done because data "vvas not available to determine the exact distribution of the super- visor's time, but it was evident that she gave the major attention to the grades measured. Then, too, she was better qualified to helji (he teachers of tlie upper grades. APPENDIX STATISTICAL DATA AlVD TABLES USED W COMPUTATION TABLE I— PROGRESS IN READING Grades IV V VI VII Average - Decimal Part of the Normal Work of School Grade Completed in Five Months. Supervised Control Per cent by Which Progress in Supervised Group Ex- ceeded Progress in the Control Group. 33% 500% 150% 500% 152% TABLE II— PROGRESS IN SPELLING Grades Decimal Part of the Normal Work of a School Grade Completed in Five Months. Per cent by Which Progress in Supervised Group Ex- ceeded Progress in the Control Group. Supervised Control IV V VI VII .5 1.0* .7* 1.5* .3 .1 .2 .7* 67% 900% 250% 114% Average .93* .33 182% TABLE III— PROGRESS IN FOUR FUNDAMENTALS Grades Decimal Part of the Normal Work of a Scnool Grade Completed in Five Montns. Per cent by Whicn Progress in Supervised Group Ex- ceeded Progress in the Control Supervised Control Group. IV V VI VII .7* .8* 1.8* 1.1* .5 .1 .5 .3 40% 700% 160% 267% Average .98' .,37 1 16.";% TABLE IV— PROGRESS IN REASONING IN ARITHMETIC Grades Decimal Part of the Normal Work of a School Grade Completed in Five Months. Per cent by Which Progress in Supervised Group Ex- Supervised Control ceeded Progress in the Control Group. IV V VI VII .9' l.I* .8* .8* .7* .4 ■8* .4 29% 175% 0% 100% Average .90* .58 55% The asterisk (*) is used to indicate when progress is normal or above. 36 Valxje op SrPERvisiox TABLE V^MEASURES OP VARIABILITY (GRADES IV— VII) Fall Subjects Supervised Group; Reading Spelling Arith.— Comp._. Arith. — Reas. ._ Control Group: Reading Spelling Arith. — Comp... Arith. — Reas Number Tested Spring 164 41.12 160 24.12 166 37.43 147 49.72 167 42.39 167 25.73 164 37.24 148 49.44 S. D. Dist. Average 7.48 45.00 7.67 28.56 9.94 44.90 12.28 58.80 6.95 43.27 7.75 27.71 8.31 41.19 11.58 .55.36 8.49 8.59 11.78 13.42 6.21 8.06 8.57 12.18 1. The standard deviation of the distribution (S. D. Dist.) is used to picture roughly the concentration of the scores about the average. 2. A distance of S. D. above and below the average includes approx- imately the middle two-thirds of the scores. 3. A range of six times the S. D. Dist. includes the bulk of the scores. TABLE VI— MEASURES OF RELIABILITY (GRADES IV— VII) Subjects Supervised Group: Reading SpelHng Aritti. Comp Aritti. Reas Control Group: Reading Spelling Aritn. Comp Arith. Reas Fall Spring Number Tested Average S. D. Av. Average S. D. Av. 164 41.12 .58 45.00 .58 160 24.21 .60 28.56 .71 166 37.43 .77 44.90 .91 147 49.70 1.01 58.80 1.10 167 42.39 .53 43.27 .48 167 25.73 .60 27.71 .62 164 37.24 .65 41.19 .67 148 49.44 .95 55.36 1.01 1. The standard deviation of the average (S. D. Av.) is used to test the reliability of the averages obtained in this experiment. 2. In interpretation it means that the chances are 365 to 1 that the true average of any large group, of which this group was a random sample, is between the averages obtained in this experiment and plus or minus 3 S. D. Av. In Consolidated Schools 37 TABLE VII— MEASURES OF RELIABILITY (CONT.) Subjects Average Improvement Difference S. D. Diff. Experimental Supervised Control Coefficient 3.88 4.35 7.47 9.08 .88 1.98 3.95 5.92 3.00 2.37 3.52 3.16 .72 .94 1.13 1.49 1.44 Spelling Arith. Comp Arith. Reas .91 1.10 .76 1. An experimental coefficient of 1.0 represents practical certainty that if the experiment were repeated the differences would favor the supervised group. TABLE VIII— STATEMENT OP CHANCES THAT IN A SIMILAR EXPERIMENT THE DIFFERENCES WOULD FAVOR THE SUPERVISED GROUP Subjects Reading Spelling Arith. Comp. Arith. Reas.. Exper. Coefficient Approx. Chances 1.44 .91 1.10 .76 20,000 to 1 160 to 1 930 to 1 60 to 1 38 Value of Supeevision TABLE IX— GRADE SCORE TABLES Age in Months Grade Score Th.-McCall Reading T Score Woody, Series B Arith. Buckingham Reasoning Crude Score Mor.-McCall Spelling No. Correct Nat'l Intel. Scale A Crude Sc. 108 3.0 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4.0 4.1 4.2 4 3 4.4 4.5 4.6 4.7 4.8 4 9 5.0 5.1 5.2 5 3 5,4 5.5 5.6 5.7 5.8 5.9 6.0 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 7.0 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 8.0 8.1 8.2 8.3 8.4 8.5 8.6 8.7 8.8 8.9 16 18 20 21 23 25 28 30 31 32 33 34 35 36 37 38 39 40 33 34 37-39 110 ___ 32 15 40-42 43-45 112 33 36 16 17 18 46-47 48-49 34 50-51 113 38 52-53 35 36 19 54-56 115.... 40 42 44 46 48 57-59 20 60-61 116 37 62-63 21 22 23 64-65 120 38 39 66-67 122 68-69 70-77 124 40 24 72 73-74 126 41 50 25 75-76 77-78 128 42 41 42 52 26 79 81-82 43 54 27 28 82-83 131 43 84-85 44 45 56 86-87 134 . 44 29 88-89 58 90-91 136 46 45 46 47 48 49 50 30 90-93 60 94 138 ... 47 31 95-96 62 97-98 141 48 49 32 99-100 64 101-102 144 33 103-104 50 51 52 53 66 105 146 ... 34 106-107 51 68 108 150 35 109 52 53 55 54 54 70 110 153 36 111-112 72 113 156 37 114-115 56 74 116-117 15(9 55 118-119 76 38 120 163 . . 56 57 121-122 39 123 166 78 124-12^ 57 58 40 126-127 170 80 128-129 59 130-131 173.... 58 82 41 132 60 135 176 134-135 59 84 136 180 61 42 137-138 139 185 62 86 140-141 142 189 60 63 43 143 1 88 1 144 « UNIVERSITY OF N.C. AT CHAPEL HILL 00034026401 FOR USE ONLY IN THE NORTH CAROLINA COLLECTION