CmttitxutinnB lu Ctoration 
 
 'fH 
 
Bonk- n?4-4 
 CpBiigk N? 
 
 COn/MCHT DEPOSIT. 
 
Factors Controlling Attendance 
 In Rural Schools 
 
 By 
 GEORGE H. REAVIS, Ph. D. 
 
 Teachers College, Columbia University 
 Contributions to Education, No. 108 
 
 Published by 
 
 reachers College, Columbia University 
 
 New York City 
 
 1920 
 
 " noera'^fe 
 
Copyright, 1920, by George H. Reavis 
 
 
 AUG 15 1921 
 g)C!.A6305^6 
 
ACKNOWLEDGMENTS 
 
 For any merit this study may possess, grateful 
 acknowledgment is here made to Professor George 
 D. Strayer, under whose direction the problem was 
 defined, and to Professor E. L. Thorndike, under 
 whose constructive criticism the treatment of the 
 material was developed. 
 
 G. H. R. 
 
CONTENTS 
 
 I. The Problem 1 
 
 11. Factors Influencing Attendance 9 
 
 III. Administrative Policies 17 
 
 IV. The Material and Method Used 26 
 
 V. Preliminary Survey of the Data 36 
 
 VI. Analysis of Significant Factors 52 
 
CHAPTER I 
 THE PROBLEM 
 
 For one who seeks to determine the factors which influence 
 rural school attendance, a number of interesting problems pre- 
 sent themselves. Do country children go to school because they 
 live within a mile, rather than three miles, of the school ; 
 because some means of transportation is provided; or because 
 the road is macadam or gravel rather than dirt? Do they go 
 because they are nine, ten, or eleven years old, rather than six 
 or sixteen; or because they are ahead in school, rather than 
 behind for their ages? Or do they go on account of the num- 
 ber of children in school, or because of the games they play? 
 
 Does the kind of teacher, her training, salary, experience, and 
 her success in teaching have anything to do with the attendance 
 of pupils? Is there any relation between attendance and the 
 amount of money invested in the school property; or between 
 attendance and the kind of school building, its size, lighting, 
 heating, and ventilation ; or between attendance and the amount 
 of equipment and teaching supplies? To what extent is the char- 
 acter of the school a determining factor in rural school attend- 
 ance? Is there any relation between attendance and the wealth, 
 general intelligence, and permanent educational interest of the 
 school patrons? 
 
 In short, what are the factors that influence the school attend- 
 ance of country children? How do the several factors operate 
 when taken together? How does each factor influence attend- 
 ance when all the others are eliminated? These are the chief 
 questions considered in this study. 
 
 The Distance Factor in State Legislation 
 
 Our lawmakers have dift'ered in regard to the effect of dis- 
 tance on school attendance. The compulsory attendance laws 
 of twenty-two states require children to attend school regard- 
 less of the distance they live from school, or of the means pro- 
 
2 Factors Controlling Rural School Attendance 
 
 vided for their transportation; the laws of thirteen states 
 exempt pupils living at specified distances from the school, 
 unless they have free transportation ; and the laws of the 
 remaining thirteen states exempt children living beyond specified 
 distances regardless of the means available for their transpor- 
 tation. 
 
 The states in each of these groups are widely distributed over 
 the whole country. Among those making no exemption for dis- 
 tance are New York, Illinois, Arizona, Idaho, and Washington ; 
 among those exempting children at certain distances, unless 
 transportation is provided, are Virginia, Iowa, Wisconsin, Ne- 
 1)raska, and North Dakota. And among the states exempting 
 children beyond a specified distance, regardless of any means 
 for their transportation, are Missouri, Arkansas, South Dakota, 
 Utah, Wyoming, Oregon, and California. 
 
 As transportation is seldom provided for children in one- 
 teacher country schools, the compulsory attendance laws in 
 twenty-six states excuse a large number of children living at 
 a distance from the schoolhouses. In most instances the dis- 
 tance at which children are excused is from two to two and 
 one-half miles. In Georgia and New Mexico, it is three miles. 
 In four states the distance at which children are excused is 
 not specifically stated in the law ; in Minnesota, a child is not 
 required to attend unless there is a school within "a reasonable 
 distance" ; in Nevada, a child is excused from the attendance 
 law if his residence is at such distance from the school as to 
 render attendance "impracticable" ; in South Dakota, the law 
 excuses those whose attendance, if required, would not be 
 "humane" ; and in Wyoming, the attendance law does not apply 
 to pupils on whom "the provisions of this act might work a 
 hardship." In Michigan, the exemption for distance applies 
 only to children under nine years of age. In Oregon, children 
 nine and ten years old, without free transportation, are excused 
 if they are more than a mile and a half from school ; older chil- 
 dren are not excused unless they are without free transportation 
 and live more than three miles from school. 
 
 Similar differences have characterized legislation on school 
 consolidation. In four states, transportation has bee required 
 
The Problem 3 
 
 for all pupils, when districts are consolidated, regardless of the 
 distances the pupils live from the school. This is the law in 
 Alabama, Iowa, Nevada, and Rhode Island. In fourteen states, 
 transportation has been required for all living- beyond a specified 
 distance. This practice is found in Ohio, Indiana, Wisconsin, 
 and Texas. In eleven states, of which New Jersey, Nebraska, 
 Colorado, and Washing-ton are examples, the boards of consoli- 
 dated districts are authorised to transport children living beyond 
 certain distances. In four states, Kentucky, Idaho, Arizona, 
 and Oregon, the laws require the approval of the voters before 
 the school boards of consolidated districts can transport chil- 
 dren at public expense; and in seven states, of which North 
 Carolina, New Mexico, and Wyoming are examples, districts 
 may be enlarged and consolidated, but it is not lawful, even by 
 a vote of the people, to transport any children at public expense. 
 
 The Pennsylvania law^ provides that every child in an aban- 
 doned district shall be furnished transportation if he lives one 
 and one-half miles, or more, from the consolidated school, 
 regardless of how far he had to walk before ; but no limit is 
 set for children in the district which was not abandoned in form- 
 ing the consolidation. 
 
 In Missouri, transportation must be provided for children 
 living more than two and one-half miles from the school, unless 
 they lived more than two and one-half miles from their schools 
 before consolidation. Children who lived more than two and 
 one-half miles from school before consolidation can be lawfully 
 denied transportation at any distance after consolidation. 
 
 In the absence of definite state regulation, the practice within 
 a state is usually more varied than the dififerences between 
 states. In discussing distances and transportation in Massa- 
 chusetts in 1898, the agent of the State Board of Education 
 said :" "There seems to be no concensus of opinion regarding 
 what is a reasonable walking distance. . . . It is the one 
 difficult question." A reasonable walking distance is ^s much a 
 problem now as it was in 1898. 
 
 1 The School Code, Harrisburg, 1915, p. 81. 
 
 - Fletcher, G. T., Consolidation of Schools and the Conveyance of Children, 
 Massachusetts State Board of Education, Boston, 1898, p. 15. 
 
4 Factors Controlling Rural School Attendance 
 
 The Problem in Maryland 
 
 The schools of Maryland are supported jointly by the county 
 and the state, and are controlled by the county board of educa- 
 tion, which has authority to change attendance district bound- 
 aries, org-anize new schools, and close old ones. The law pro- 
 vides that in closing a school, no pupil shall be required to go 
 an unreasonable distance unless transportation is provided. The 
 State Superintendent of Schools is the interpreter of the school 
 laws and "shall decide, without expense to the parties concerned, 
 all controversies and disputes, involving the proper administra- 
 tion of the public school system." ^ As a result, the State De- 
 partment of Education is sometimes called upon to rule on the 
 "true intent and meaning" of "an unreasonable walking dis- 
 tance." 
 
 A reasonable distance over a state road might be an unreason- 
 able distance over an unimproved road. A reasonable distance 
 over an unimproved road in the sandy soil of Southeastern 
 Maryland might be an unreasonable distance on the other side 
 of the Bay in a mud-producing clay soil; and a reasonable dis- 
 tance in either of these places might be an unreasonable distance 
 if it lay across ravines in the mountains of Western Maryland. 
 
 Some county boards of education, anxious to run the schools 
 as economically and efficiently as they can, accept every oppor- 
 tunity to close a small school and enlarge those that remain, so 
 that the same total expenditure will permit them to pay more 
 per teacher and to secure a higher grade of service. Some 
 other county boards, more anxious to satisfy the local demands 
 of each small group of patrons, conduct more small schools with 
 a correspondingly lower salary schedule for the same total ex- 
 penditure. We then face such questions as, how does the size 
 of the district aflfect attendance? To what extent does progress 
 in school depend upon attendance? How much better should 
 the teacher, building, and equipment be to justify a county 
 board of education in placing the school farther from the homes 
 of the children ? How large should a country school attendance 
 district be? How small a school should the state permit? 
 
 '^Maryland School Law (1918), p. IS. 
 
The Problem 5 
 
 The size of a country school district is limited by the dis- 
 tance children will, under the different conditions, go to school. 
 Before these questions can be answered, the factors influencing 
 the attendance of country children must be isolated, analyzed, 
 and correlated with progress in school. 
 
 The improvement of the educational advantages for any group 
 of children involves two variables: the number of days' attend- 
 ance, and the quality of the school training. Improvement is 
 made by improving either, and the maximum improvement 
 comes when both the attendance of the children and the quality 
 of their school training are at the best. We have been accus- 
 tomed to lay much more stress upon the quality of the school- 
 ing, with a minimum of attention to the factors determining 
 the amount of exposure children get to the instruction offered. 
 Improvement in the quality of the schooling offered might or 
 might not be directly related to the holding or drawing power 
 of the school. No doubt enlarging school districts and paying 
 more for teachers give a higher type of school training for 
 the children who attend; and if the kind of school has more to 
 do with attendance than the distance children live from it, the 
 consolidation of small districts into larger ones has an addi- 
 tional argument in its favor. 
 
 It is not the purpose of this study to determine what changes 
 in the rural school will improve the quality of the schooling 
 offered country children. It merely attempts to isolate and 
 analyze the factors that influence the school attendance of coun- 
 try children, in order that, in so far as the factors can be con- 
 trolled, we may better understand how to recast the school situa- 
 tion to increase the number of days of school training country 
 children receive. 
 
 Related Studies 
 
 Burnham,^ in his study of Tzvo Types of Rural Schools, and 
 Miller,^ in his Rural Schools in Canada, decided that the 
 attendance of country children is better in consolidated districts 
 
 ^Burnham, Ernest, Teachers College, Columbia University, Contributions to Educa- 
 tion, No. 51. 
 
 '■= Miller, J. C, Teachers College, Columbia University, Contributions to Edu- 
 cation, No. 61. 
 
6 Factors Controlling Rural School Attendance 
 
 with free transportation than in one-teacher schools without 
 transportation. In a bulletin on Consolidated Rural Schools, 
 published by the United States Department of Agriculture in 
 1910, Knoor^ reached a similar conclusion. But there has 
 been no comprehensive study of school attendance, in either 
 rural or urban communities, and the studies that have been made 
 of the persistence and progress of pupils in city schools, not 
 being directly concerned with attendance, have thrown little light 
 on the relation of attendance to progress in school, or on the 
 factors controlling attendance. 
 
 The pioneer studies of retardation and elimination by Thorn- 
 dike- and Ayres^ showed that the retarded pupil is much more 
 often eliminated, but neither had the data to determine the rela- 
 tion between attendance and retardation. In his study Prog- 
 ress Through the Grades of City Schools,'^ Keyes notes that 
 a child who misses a certain number of days more often fails 
 of promotion than does a child whose attendance is regular ; and 
 the Committee on School Inquiry^ in New York City found 
 a definite relation between attendance and promotion ; but neither 
 study established the fact that the poor attendance was the cause 
 of the failure to be promoted, or the fact that inferior school 
 work and lack of interest was the cause of the poor attendance. 
 It has been assumed frequently that attendance in itself is not an 
 important cause of retardation in city schools. "On the whole, 
 the effect of absence is small until very large amounts of absence 
 are reached."*' But we have few facts on which to base an opin- 
 ion, and it would not necessarily apply to rural school condi- 
 tions, even if the opinion held for city schools. 
 
 Holley'^ investigated The Relationship Betzveen Persistence in 
 School and Home Conditions in six small cities of Illinois, and 
 came to the conclusion that "there is a close relationship between 
 the advantages of a home, its educational, economic, and social 
 
 1 Bulletin No. 232, United States Department of Agriculture. 
 
 =• Bulletin No. 4, 1907, United States Bureau of Education. 
 
 ' Ayres, Leonard P., Laggards in Our Schools, New York, 1909. 
 
 * Keyes, C H., Teachers College, Columbia University, Contributions to Education, 
 No. 42. 
 
 » Report of the Committee on School Inquiry, New York City. 
 
 ' Strayer, G. D. and Thorndike, E. L., Edticaiional Administration, Macmillan 
 Company, 1914, p. 42. 
 
 ' Hollev, C. E.. Fourteenth Year Book, National Society for the Study of Edu- 
 cation, 1914, pp. 96. 100. 
 
Tile Problem 7 
 
 stations, and the number of years of schooling which its chil- 
 dren receive" and "if a person wished to forecast from a single 
 objective measure, the probable educational opportunities which 
 the children of a home have, the best measure would be the 
 number of books in the home." He also thinks that early elimi- 
 nation is "largely due to factors outside the school," but his 
 study did not include any factors inside the school. 
 
 Van Denburg^ sought "to determine by a large series of 
 carefully collected measurements, the kind of pupils who leave, 
 as compared with the kind who stay" in New York City high 
 schools, and found that "early elimination from high school is 
 favored by a late entering age ; by having younger brothers or 
 sisters ; by a childhood free from serious illness ; by foreign- 
 born parentage of Irish, Austro-Hungarian, Scotch or Italian 
 stock; by the choice of business as an occupation by boys or 
 stenography by girls ; by a disbelief in the value of a high school 
 course; by an uncertainty as to probable length of stay or a 
 determination to leave early," but that "on the whole, the eco- 
 nomic status of these pupils (so far as it is shown by monthly 
 rental) seems to be only a slight factor in the determination of 
 length of stay in the high schools." 
 
 In a recent and thoughtful study of failures in eight high 
 schools of New York and New Jersey, OBrien- says, "per- 
 haps one of the simplest factors with a prognostic value on 
 failure may be found in the facts of attendance," but "it hap- 
 pened, unfortunately, that the reports for attendance were in- 
 complete or lacking in a considerable portion of the records" 
 although the schools included in the study were selected chiefly 
 on account of the completeness of their scholarship records and 
 "more dependable school records than those employed are not 
 likely to be found." 
 
 Briggs,^ in his chapter on Secondary Education in the 1918 
 report of the United States Commissioner of Education, com- 
 mends the recent published reports of high school principals, 
 
 1 Van Denburg, J. K., The Elimination of Pupils from Public Secondary Schools, 
 Teachers College, Columbia University, Contributions to Education, No. 47, pp. 4, 
 113, 114. 
 
 ' OBrien, F. P., The High School Failures, Teachers College, Columbia University, 
 Contributions to Education, No. 102, pp. 29 and 11. 
 
 * Briggs, T. H., Bulletin No. 47, 1918, United States Bureau of Education. 
 
S Factors Controlling Rural School Attendance 
 
 citing numerous examples of improvement in compiling and 
 using the facts pertaining to the classification and progress 
 of pupils. In regard to the facts of attendance, however, he 
 finds "several schools report the numher of pupils who were 
 present for difl^erent i)ortions of a semester, but no one of them 
 correlates the data with the facts concerning success in school 
 work, and not one of them presents any program for improving 
 the attendance." 
 
 ISOL-'\TION OF THE FACTORS 
 
 The studies of persistence and progress of school children 
 have made no attempt to isolate the factors influencing persist- 
 ence and progress so that we might know how^ any one factor 
 operates apart from its connection with others. The present 
 study attempts to isolate the factors in order to determine how 
 each factor affects attendance when taken alone. 
 
 Summary 
 
 1. This study attempts to analyze the factors that influence the 
 attendance of country school children. 
 
 2. Distance from school has been recognized in state legislation 
 as a factor influencing rural school attendance. 
 
 3. The studies of persistence and progress of school children, not 
 being directly concerned with attendance, have thrown little 
 light on the factors that influence attendance, or on what rela- 
 tion attendance bears to progress in school. 
 
 4. This study attempts to isolate the factors in order to deter- 
 mine how they operate when taken together, and separately. 
 
CHAPTER II 
 
 FACTORS INFLUENCING ATTENDANCE 
 
 This chapter reports the results of the study and summarizes 
 the factors influencing' attendance in the rural schools included 
 in the analysis. The evidence on which the statements are based 
 will be found in the chapters that follow. 
 
 An examination of more than fifty possible factors shows that 
 many of the facts analyzed have no significant relation to attend- 
 ance in the five counties included in this study. But whether 
 children live near the school, whether they are ahead or behind 
 the proper grades for their ages, whether they are at the head 
 or foot of their classes, whether they have the better paid and 
 higher-rated teachers, and whether they live in a community 
 interested in education, are important factors in the school 
 attendance of country children. 
 
 Sex of the children makes very little difference. Boys and 
 girls attend equally well, except between the ages of 12 and 14 
 years when girls go to school more days than boys. 
 
 Although the total school training of the teacher has a slight 
 relation to the attendance of her pupils, the connection is not 
 marked and is not so good an indication of how a teacher's 
 pupils will attend as is her salary, or her rating by the county 
 superintendent. The amount of professional training, consid- 
 ered apart from a teacher's total trainings, seems to have noth- 
 ing to do with attendance. 
 
 The number of years' experience of the teacher is not a con- 
 siderable factor; nor is her organization of the day's work as 
 expressed by the number of recitations in the daily program; 
 nor is the neatness of the teacher as shown by the appearance 
 of her written report ; nor is her accuracy and attention to detail 
 as exhibited by the completeness with which she makes out an 
 extended report ; nor is the quality of her handwriting. 
 
 9 
 
10 factors Controlling Rural School Attendance 
 
 The attendance in the 200 country schools included in this 
 study is not influenced by the school building nor by the equip- 
 ment, material and supplies with which the teacher works. It 
 is apparently not influenced by the number of books in the 
 library^ the kind and amount of blackboard, the age of the desks, 
 the size of the schoolroom, the floor space per pupil, the number 
 of windows and their arrangement, the heating and ventilation, 
 nor by the total value of school property. 
 
 The size of the school, as measured by the total enrollment, 
 the number attending five months and the average daily attend- 
 ance, and the recreational activities of the children, as meas- 
 ured by the number and kind of games they play, have a slight 
 but unimportant relation to attendance. The appearance of 
 the schoolroom as measured by the number and kind of pictures 
 on the walls, and the play facilities as measured by the size and 
 suitability of the playground seem to have no relation to attend- 
 ance at all. 
 
 However, the influence of the narrow range of factors which do 
 show a connection with attendance is surprisingly large. When 
 one group of children is selected, according to attendance, from 
 the top and another from the bottom of a county, the first group 
 including the twelve per cent of all the children that attended 
 best, and the second group the twelve per cent that attended 
 poorest, the two groups average about the same in age ; but 
 the first group is three grades farther along in school, and 
 lives approximately a mile nearer the school than the second 
 group. Only 10% of the first group is marked inferior in school 
 work as compared with 50% of the second; and 90% of the chil- 
 dren in the first group are promoted at the end of the year as 
 compared with 23% of those in the second group. 
 
 Selecting two groups of children from a county, as above, 
 more of the first group are under better teachers, more of them 
 are found in the better buildings, and more of them are in the 
 schools located in the better communities. 
 
 The better teachers, however, are found more often in the 
 better buildings, which, in turn, are more often in the better 
 communities. It was not possible to determine from a rough 
 analysis whether any one factor, — the teacher, the building, or 
 
Factors Influencing Attendance 11 
 
 the community, apart from the other two, is responsible for the 
 connection of the three with attendance. Again, if children liv- 
 ing farther from school are farther behind for their ages, do in- 
 ferior work, and attend fewer days, it is important to know the 
 responsibility of each of these factors separately. Do the chil- 
 dren living farther away attend fewer days because they are 
 behind in their school work, or because they have longer dis- 
 tances to come? Are they doing a poorer quality of work be- 
 cause they are behind the proper grades for their ages, or be- 
 cause they attend fewer days, or because they live farther away 
 from the school? It was necessary to use a finer method of 
 analysis to interpret the data. 
 
 There is a statistical method for untangling such factors. For 
 example, two men may contest with each other in shooting, and 
 one of them consistently may make a better score, when the dif- 
 ference is not due to a difference in marksmanship, but to the 
 difiference between the guns they are using; or they may make 
 the same score when one of them is a better marksman, the dif- 
 ference in the guns bringing the two scores together. A diffi- 
 culty such as this, however, can be easily remedied by giving the 
 contestants the same gun. But a complex group of factors as 
 they exist in nature cannot often be treated in this way. It is 
 necessary to have a method of making the scores comparable 
 when the men cannot be given the same gun. 
 
 The other way in which this can be done is to determine the 
 exact difiference between the guns, and to correct the scores to 
 allow for the difference. A correction can be made in this way 
 for any factor in which the men differ, provided the difference 
 between the two men in regard to the factor can be measured. 
 Having corrected for the difference in the guns, a similar cor- 
 rection can be made for the difference in the kind of ammunition 
 they are using, the difference in the adjustment of the lenses to 
 their eyes if they wear glasses, the difference due to excessive 
 fatigue and temporary nervousness of either, and so on, for any 
 factor that we desire to eliminate. In this way their relative 
 marksmanship can be computed for a wide range of conditions, 
 from one complete set of scores. We can determine how they 
 would shoot if they had the same gun and the same ammunition. 
 
12 Factors Controlling Rural School Attendance 
 
 if their glasses fitted them equally well, if they were in equally 
 good physical condition for the match, and so on. 
 
 Applying the method to the 5315 children in the 200 country 
 schools of five counties, the factors influencing attendance can 
 be isolated or thrown into any desired combination. One factor 
 after another can be eliminated until the desired combination is 
 obtained. Take a group of children, some of whom are farther 
 behind the proper grades for their ages, live farther from school, 
 do inferior work, and attend fewer days. We can determine the 
 kind of school work they would do if they all attend the same 
 number of days or live at the same distance; or the quality of 
 work they would do if they were equally far along in school; or 
 how far along they would have been had they lived the same 
 distance from school ; or we can determine the relation of these 
 factors in any other combination.^ 
 
 Although the better teachers were more often in the better 
 buildings, which, in turn, were more often in the better commu- 
 nities, we can compute how attendance would be related to the 
 kind of teacher if all the teachers were in the same kind of 
 buildings and communities ; if their pupils came the same dis- 
 tance to school, had reached the same grades for their ages, and 
 were equally successful in doing their school tasks. Likewise, 
 the buildings, grounds, and equipment can be isolated to deter- 
 mine how they would be related to the attendance of the chil- 
 dren if they had the same kind of teachers, were in similar com- 
 munities, and so on. 
 
 When the important factors in school attendance are freed 
 from their interrelations and dependence upon each other, we 
 find that children living farther from school attend fewer days 
 regardless of their ages for their grades, the quality of work 
 they do, the kind of teachers they have, the kind of buildings, 
 grounds and equipment, or the educational interest of the com- 
 munities in which they live. In the absence of transportation, 
 distance is the strongest single factor influencing the attendance 
 of country children. It decreases with each age-group, the corre- 
 
 1 This explanation of partial correlation, as well as the illustration of the gun 
 preceding it, is used for the sake of brevity and clearnes<: But partial correlatioa 
 is selective rather than dynamic, as explained in Chapter IV. 
 
Factors Influencing Attendance 13 
 
 lation being about .60 for 5-7-year-old children, .45 for 8-11, and 
 .25 for 12 and over. Children living more than two miles from 
 school attend only half as many days as children living within 
 a quarter of a mile of school. 
 
 Likewise, children behind the proper grades for their ages 
 attend fewer days regardless of the kind of work they do, the 
 kind of teachers and school plants they have, or the financial 
 status and educational interest of their communities. This rela- 
 tion is slightly more marked for the older age-groups. The 
 correlation betwen the age-grade relation and attendance is 
 about .30 for 5-7 years, and .40 for the two older groups. 
 
 Children attending fewer days do an inferior quality of work, 
 ranking below the pupils with whom they recite, regardless of 
 the distance they live, the grades they have reached, the kind 
 of teachers they have, or any other influence. The relation be- 
 tween attendance and quality of work is about .35 for the 5-7- 
 year group, and drops to about .25 and .20 with the older groups. 
 
 The better teachers have better attendance regardless of any 
 other factors, and the community rated higher for wealth, gen- 
 eral intelligence, and educational interest has better attendance ; 
 but the kind of building, equipment, and grounds, when freed 
 from its connection with the kind of teacher and community, 
 has no relation to school attendance. The efficiency of the 
 teacher and the kind of community are not so important in school 
 attendance as other factors, the correlation for kind of teacher 
 and attendance being about .13, and for community and attend- 
 ance about .10. The kind of teacher has a little more to do 
 with the attendance of older than of younger children. The 
 better teachers classify their pupils a little lower for their ages, 
 and mark younger pupils a little higher in quality of work and 
 older pupils a little lower than do the poorer teachers. 
 
 Thus, the distance children live from school, the progress they 
 make through the grades, and their success in doing school tasks 
 are controlling factors in the attendance of country children. 
 
 Although children living farther from school are farther be- 
 hind for their ages and do inferior work, distance is only an 
 indirect cause of retardation. Greater distance from school 
 
14 Factors Controlling Rural School Attendance 
 
 causes more absence, more absence lower quality of work ; lower 
 quality of work increases failures, and failures retard the prog- 
 ress of children through the grades. These factors have a 
 cumulative effect. Children living greater distances attend 
 fewer days, do inferior work and get farther behind, and then 
 being both farther from school and farther behind in school, 
 they lose still more time, do more inferior work, and still more 
 often fail, and get even farther behind. 
 
 Attendance is the most important determining factor of qual- 
 ity of work; and quality of work, year after year, is the deter- 
 mining factor of the grade a child reaches by a given age. 
 Being up to grade for one's age and successful in school tasks 
 also has a cumulative effect. The child who is successful in 
 his work gets promoted, keeps up with his grade, and being in 
 the proper grade for his age, attends more days, which helps 
 his quality of work and tends to keep him up to grade. 
 
 An improved road no doubt facilitates transportation, but it 
 does not greatly help the attendance of children who walk. It 
 is perhaps the lengtli of exposure in going from home to school, 
 rather than the effort of walking. A child would walk a given 
 distance over a dirt road almost as quickly as over an improved 
 road. As measured by its influence on the attendance of chil- 
 dren who walk, a mile of macadam road is equal to three- 
 fourths of a mile of unimproved dirt road in the five counties 
 included in this study. But transportation has a much more 
 marked effect. The provision of some means of regular trans- 
 portation seems almost to eliminate the distance factor. 
 
 Some Implications of the Findings 
 
 The assumption in the laws of a number of states that dis- 
 tance from school becomes a serious handicap at about two and 
 one-half miles is erroneous. In the absence of transportation, 
 additional distance up to about three-fourths of a mile is a most 
 significant factor in attendance, but additional distance beyond 
 one mile does not have so great cumulative effect. While chil- 
 dren living within a fourth of a mile attend much better than 
 those living farther away, additional distance up to a mile and a 
 
Factors Influencing Attendance 15 
 
 half has no marked cumulative effect. A child living within a 
 mile and a half attends almost as well as one living within a 
 mile or within a half mile, but children living beyond a mile 
 and a half show a marked drop in attendance. There seems to 
 be three rather distinct zones with somewhat similar attendance : 
 children living within a quarter of a mile; those living more 
 than a quarter of a mile but not more than a mile and a half ; 
 and those living more than a mile and a half from the school. 
 If all children living beyond a given distance are to be trans- 
 ported, that distance should be one-fourth of a mile rather than 
 a mile or more. In the absence of some means of conveyance, 
 country children living beyond a quarter of a mile from the 
 school do not have equal educational opportunity with those 
 living nearer the school. 
 
 The problem of improving the attendance of country school 
 children is not solved simply by improving the quality of the 
 school. The pulling power of new and modern buildings, addi- 
 tional equipment, and larger and more suitable school grounds, 
 loses out entirely in competition with the more potent factors 
 of distance, age-grade relation, and success in school work. An 
 improvement of the teacher is accompanied by an improvement 
 in the attendance, but this is perhaps due chiefly to the fact that 
 the teachers rated higher more often interest their pupils in 
 the school tasks. A wide range of measures of the teacher 
 that ought to be some measure of ability and efficiency in teach- 
 ing has practically no relation to attendance. The kind of 
 teacher is so small a factor in school attendance that attendance 
 will not be greatly helped by improving the teacher alone. Im- 
 proving the educational facilities of country children presents 
 two distinct and almost separate problems : improving the qual- 
 ity of the schooling offered, and increasing the attendance of 
 the children for whom the schooling is provided. The problem 
 has not been solved when good schools have been provided. 
 The regular and continued attendance of the children must be 
 secured. 
 
 No data were available to check the influence of an active inter- 
 est in attendance on the part of the teacher. While the training 
 and rating of the teacher do not bear a very important relation 
 
16 Factors Controlling Rural School Attendance 
 
 to attendance, it should be remembered that no evidence has been 
 presented showing how much the teacher might improve attend- 
 ance by personal interest and solicitation. Nor should the data 
 be construed to minimize the importance of good school equip- 
 ment and buildings. The evidence, however, does show that 
 these things have little to do with school attendance. 
 
 The close relation between attendance, success in school work, 
 and progress in school, together with the cumulative effect of 
 attendance and absence, emphasizes the importance and economy 
 of regular attendance. 
 
 There is good evidence that the deep-seated educational in- 
 terest of a community is only slightly less important as a factor 
 in school attendance than the kind of teacher. Since such in- 
 terest, no doubt, has a cumulative effect, the attitude of the 
 patrons toward the school might w-ell receive more attention and 
 encouragement. 
 
 Since the distance factor reduces attendance by half for chil- 
 dren living two miles from school, a large number of country 
 children living remote from school do not have equal educa- 
 tional opportunity with other children. The next chapter will 
 consider this and some other administrative problems in con- 
 nection with rural school attendance. 
 
CHAPTER III 
 
 ADMINISTRATIVE POLICIES WITH REFERENCE TO 
 RURAL SCHOOL ATTENDANCE 
 
 A pupil's progress in a rural school seems to be a function 
 of his attendance. His standing in the class with which he re- 
 cites has a definite and fixed relation to the number of days he 
 is present. The quality of work he does determines his chance 
 of promotion. His promotion determines the grade he will 
 reach by a given age. The grade he reaches by a certain age 
 again has a marked efifect upon his attendance. Thus attendance 
 in any year, by afifecting a pupil's class standing, also effects his 
 future attendance. The pupil who is irregular in attendance and 
 drops in the quality of his work, attends fewer days the follow- 
 ing years as a result of being behind ; while the regular pupil 
 who keeps up to standard accumulates a momentum that helps 
 to carry him on. The child who is successful in his work gets 
 promoted, keeps up with his grade, and being in the proper 
 grade for his age, attends more days, which helps his quality 
 of work and tends to keep him up to grade. Thus school attend- 
 ance is cumulative in its effect. 
 
 The relation of attendance to progress in school, and the 
 cumulative aspect of attendance, emphasize the importance of 
 regular attendance in the rural schools. Much of our present 
 investment in rural schools is rendered ineffective by the failure 
 of pupils to attend. 
 
 Although the counties included in this study have full-time 
 attendance officers, with their traveling expenses paid, the pres- 
 ent machinery for supervising school attendance is not adequate 
 for the task. It is probable that a program of law enforcement 
 alone cannot fully meet the need.^ But closer supervision of 
 
 ^ The attendance officer in Baltimore County is very active and is considered 
 very efficient. Althongh prosecution has not been the main feature of his effort, he 
 made 122 arrests in the year 1916-1917, brought 108 of the cases to trial and did not 
 lose a case. (See Annual Report, Maryland State Board of Education, 1917. p. 48.) 
 But the attendance in the rural schools of Baltimore County is not markedly better 
 than in the other counties. It is not so good as the attendance in Queen Anne 
 County, which has more state roads and a larger proportion of tiie children trans- 
 ported. 
 
 17 
 
18 Factors Controlling Rural School Attendance 
 
 attendance in rural schools should 1)c a part of the pro£i;ram of 
 every state department of education. The term "truant officer" 
 in the majority of our state laws and in much of our educa- 
 cational literature ' indicates the view too often taken of the 
 problem of securing satisfactory school attendance. An attend- 
 ance officer should understand the chief causes of absenteeism, 
 and, with the co-operation of the teachers and principals, pro- 
 mote the regular attendance of all the pupils. 
 
 In the light of these facts it would seem that the attendance 
 officer's task should be approached in a different manner. In 
 the majority of our state laws and in much of our educational 
 literature, an attendance officer is called a truant officer. State 
 laws providing for an attendance officer emphasize the officer 
 aspect of the work, stressing the policing function, his authority 
 to make arrests, and to bring offenders to justice. When at- 
 tendance or "truant" officers are not provided, the laws usually 
 designate the town marshals or deputy sheriffs. This tendency 
 in drafting legislation, and the terminology we use, indicate 
 the current conception of the job. Regardless of what we call 
 tlie person employed to secure regular attendance, the task is 
 larger than policing the territory for absent children during 
 school hours. Wq should conceive the duties of the position 
 more as a "super\isor of school attendance." A supervisor of 
 school attendance should be able to analyze the conditions sur- 
 rounding a child, determine the cause of the absence, and, so 
 far as it is remediable, make the necessary adjustments. Such 
 a supervisor should also be able to sit down with an uneducated 
 parent and justify the class-room methods and practices of the 
 school. Only two states, however, prescribe educational quali- 
 fications for an attendance officer and neither of these requires 
 a knowledge of school practice, or teaching experience. 
 
 In Maryland - preference has been given for the position of 
 attendance officer to women who have had training as social 
 
 ^ For use of the term "truant officer," see Bulletin No. 23, 1918, p. 139, and 
 Bulletin 47, 1915, p. 527, U. S. Bureau of Education; Betts, G. H., New Ideas in Rural 
 Schools, Riverside Educational Monographs, edited by Henry Suzzallo, p. 22; Strayer, 
 G. D., City School Expenditures, Teachers College, Columbia University, pp. 20-27; 
 Miller, J. C, Rural Schools in Canada. Teachers College, Columbia University, p. 88; 
 Fifteenth Year Book, The National Society for the Study of Education, Part II, p. 80. 
 
 ^ See Annual Report, State Board of Education, 1916, p. 7, Teachers Year Book, 
 Maryland State Board of Education, 1916-1917, p. 37. 
 
Administrative Policies 19 
 
 workers, in the belief that a trained social worker can better 
 locate the cause back of the family's educational apathy, and the 
 absence of the children. A woman has been favored because it 
 is thought she can enter the home and investigate a case with 
 less embarrassment, as most of the interviews are with the 
 mothers. In London^ recent opinion favors combining the 
 duties of the school nurse and attendance officer with the idea 
 that attendance problems are closely related to the sanitary and 
 hygienic problems of the home. 
 
 Since an attendance officer must continually justify the policies 
 and methods of the schools to parents who do not understand 
 present-day school methods, one of the requirements might be a 
 first-hand knowledge of class-room practice. The Indiana law 
 requires schooling equivalent to the completion of the eighth 
 grade, and Alaryland requires preparation equivalent to a stand- 
 ard normal school course of two years above high school gradu- 
 ation. The qualifications of the attendance officer should be 
 prescribed in state laws in the same way in which requirements are 
 laid down for teachers' certificates. An attendance department 
 should endeavor to locate and control the factors that influence 
 the attendance of children, rather than to deal simply with the 
 chronic cases of absenteeism referred to it. Its chief function 
 might well be preventive rather than curative, similar to the 
 function of a modern department of health. 
 
 The problem of school attendance is not solved merely by 
 providing a higher type of instruction, together with adequate 
 buildings, grounds, and equipment. This does not mean that 
 we should put less money into buildings, grounds, and equip- 
 ment ; it means that a good physical plant alone is not a solu- 
 tion of the attendance problem. It does mean that additional 
 preparation and improvement in the teachers themselves will 
 not greatly improve rural school attendance. Although the 
 teacher's training, both academic and professional, the quality 
 of her handwriting, and the neatness and accuracy of her re- 
 ports, and other characteristics generally recognized as elements 
 of an efficient teacher, have no important relation to the regu- 
 
 ^ See Report of Chief Medical Officer, English Board of Education, London, 1916. 
 
^0 factors Controlling^ Rural School Attendance 
 
 larity with which her pupils attend, there is, however, some- 
 thing about the teacher which does have a fixed relation to 
 school attendance. It enters into the rating- she is given by the 
 county superintendent, and again as one of the factors deter- 
 mining her salary. It may be the intangible thing we call per- 
 sonality, a function of her attitude towards the children, and 
 composed of those personal qualities which enable her to arouse 
 the interest of her pupils in their work. Since some teachers 
 have better attendance than others, and there is evidence that 
 the difference is due to personal qualities, an added responsi- 
 bility is placed on the teacher for taking a personal interest in 
 her pupils. Each day she adds to a child's attendance is as 
 much a service to the cause of education as a corresponding 
 improvement in the quality of her teaching. Teachers should be 
 led to feel this responsibility. 
 
 Lack of interest is probably the cause of poor attendance by 
 children who are behind the proper grades for their ages. Chil- 
 dren, as well as adults, do not take an active and sustained in- 
 terest in tasks at which they continually fail. Anything that 
 gives the pupils greater interest in their work helps their attend- 
 ance. This conclusion is in line with the conclusions reached in 
 studies concerning the reasons why city children leave school. 
 Of the pupils leaving elementary schools, Wooley^ found "75% 
 of the families did not need the children's earnings. The chil- 
 dren could have continued in school had they desired.'' Similar 
 conclusions have been reached by other studies.^ Thus an 
 active and sustained interest in their school work seems to be 
 an important factor in the attendance of rural school children. 
 This should be recognized by school administrators and every 
 agency utilized to keep up such interest. 
 
 Since the wealth, general intelligence, and educational interest 
 of the school patrons, as estimated by the county superintend- 
 ents, have almost as much influence on attendance as the kind of 
 
 1 Wooley, Helen T., Facts about the Working Children of Cincinnati. 
 
 "See Conditions Under Which Children Leave School to Go to Work. Washington, 
 1910, Senate Document No. 645, p. 57, which found that 29.3 per cent needed to go 
 to work; Report of the Co>nmL<:sion on Industrial and Technical Education (in Massa- 
 chusetts), Teachers College, Columbia University, Educational Reprints. No. 1, 1905, 
 p. 92, which found that 24 per cent needed to go to work; Barrows, Alice P., Report 
 of the Vocational Guidonce Sun'ey, Bulletin No. 9, Public Education Association, 
 New York City, 1912, which found that 20 per cent needed to go to work. 
 
Administrative Policies 21 
 
 teacher, and since the community interest is perhaps cumu- 
 lative, it seems that the attitude of the rural community toward 
 its school is a matter of considerable concern to those connected 
 with the administration of state school systems. Parent-teacher 
 associations have been more extensively used in towns and cities 
 to stimulate local interest in school systems. It would seem that 
 more might be done along this line in rural communities. 
 
 But of more importance still is the size of the district and the 
 means of transportation for country children. Distance from 
 school is the strongest factor influencing the attendance of pupils 
 enrolled in the rural schools of the five Maryland counties in- 
 cluded in this study. The data for this study did not furnish 
 sufficient cases to make an exhaustive study of transportation. 
 The study is concerned with transportation only as it is neces- 
 sary to translate days transported into equivalent distances 
 walked. But there seems to be very good evidence that trans- 
 portation practically eliminates the distance factor. 
 
 School administrators, however, have often assumed that the 
 educational facilities of rural districts will be improved merely 
 by enlarging the districts. 
 
 While in city and town schools departmental work is more and 
 more stressed to develop skill and specialized effort, the farmer, in 
 many localities, seems bent upon increasing the number of schools, 
 desiring apparently to multiply the miserable little "shoe-box" 
 structures near his door rather than to have real educational insti- 
 tutions farther away.* 
 But this traditional prejudice of country people in favor of a 
 school near their homes seems, in the absence of transportation, 
 to be well founded. 
 
 A smaller school near the homes of the people it serves means 
 less expenditure per teacher and less desirable building, grounds, 
 and equipment ; larger districts permit of higher salaried teachers 
 with the same efifort, and better buildings, grounds, and equip- 
 ment; but the apostles of school consolidation have not clearly 
 recognized the distance factor in school attendance, and have not 
 squarely faced the issue. Consolidation alone will not solve the 
 problem. It is not enough to enlarge the districts. Something 
 
 1 Frittain, M. L., The Consolidation of Coutitry Schools, Georgia Campaign Com- 
 mittee, 1912, p. 6. 
 
32 Factors Controlling Rural School Attendance 
 
 must be clone to overcome the barrier of distance. It is not sound 
 policy to argue that we may have "Consolidation v^ith or v^ith- 
 out Transportation," which, for example, is the title of a bul- 
 letin of the University of Texas, saying, "When only two or 
 three schools are consolidated and when none of the children 
 are placed thereby at great distance from the school, free trans- 
 portation need not be provided."^ The traditional prejudice of 
 country people in favor of a school near their homes should be 
 respected, unless provision is made for getting their children 
 to school. Consolidation without transportation should be pro- 
 hibited. 
 
 Transportation may be impracticable in small rural districts, 
 but transportation and improved school attendance have not 
 been the arguments in favor of enlarging the districts. It is 
 probable, however, that the improved attendance resulting from 
 transportation accounts for a large share of the improvement 
 noticed when schools have been consolidated. Such data as are 
 available in the studies of Burnham,^ Miller,^ 'Knoor,- and 
 others indicate that consolidated schools have better attend- 
 ance. It is probable that they have better attendance not be- 
 cause they are consolidated schools but because transportation to 
 them is provided. 
 
 Arp, in a comprehensive summary of Rural Education and 
 the Consolidated School, finds "if children can be transported 
 safely, rapidly, and comfortably, all other objections may be 
 speedily overcome," ^ although he does not emphasize better 
 attendance as an argument for consolidation. The word attend- 
 ance does not occur in the four-page index to his book of two 
 hundred pages. His summary of the handicaps of the rural 
 school is typical of our best literature on the subject. "Three 
 great dead-weights that hang like millstones from the neck of 
 rural school progress are: (1) The lack of real professional 
 supervision; (2) the small district unit of taxation; and (3) the 
 untrained teacher."* 
 
 ^ Bulletin No. 96, 1907, University of Texas, p. 3. Imagine two or three districts 
 in Texas consolidated without placinf? any of the children at a great distance from 
 school. 
 
 ^ Already cited. 
 
 "p. 23. 
 
 * Arp I. B., Rural Education and the Consolidated School, World Book Company, 
 1918, p. 162. 
 
Administrative Policies 23 
 
 Updegraff's summary of the charge against the rural school 
 mentions attendance but does not emphasize it. "Ever siince the 
 time of James C. Carter, and of Horace Mann, educators have 
 called attention to the deficiencies of the country school. The 
 counts in the indictment which has been brought against it 
 through all these years differ but little, — unattractive sites, mis- 
 erable buildings, insufficient equipment, poorly prepared and 
 paid teachers, inadequate and incompetent supervision, unevenly 
 distributed enrollment, irregular attendance, meager curriculum, 
 and a poorly conducted school." ^ 
 
 If the distance factor is strong enough to reduce the days 
 attended one half at a distance of two miles from school, it would 
 seem that children living remote from schools, without some 
 means of transportation, do not have equal educational opportu- 
 nity with those residing nearer. States interested in equal edu- 
 cational opportunity for country children have another opportu- 
 nity here. Not only should state aid be used to supplement the 
 general resources of weak rural districts ; but something more 
 should be done for that large group of children in all country 
 schools who live at some distance from the school. A state can- 
 not place a school within a quarter of a mile of every home ; but 
 it can encourage free transportation by removing legal restric- 
 tions and supplying liberal grants of state aid for transportation. 
 
 While most of the states either authorize or require the trans- 
 portation of children in consolidated districts, the distance at 
 which transportation usually begins is too great. If children 
 are to be transported at public expense, a quarter of a mile 
 rather than two miles should be the starting point. 
 
 The present laws on transportation are far from adequate. In 
 some states it is unlawful for a district school to use public 
 funds for transportation of pupils. It is customary for the 
 law to enumerate the powers of the district board without men- 
 tioning transportation and the courts have usually held- that 
 a district board cannot transport pupils at public expense when 
 not specifically authorized to do so. 
 
 ^Updegraff, Harlan, article in Education, 41:135. 
 
 " Voorhees, H. C, The Law of the Public School System of the United States, 
 p. 83. 
 
24 r actors Controlling Rural School Attendance 
 
 From the beginning of public schools, state laws have pro- 
 vided for combining or uniting two or more districts into a single 
 large district. In this case, consolidation is as old as rural 
 school organization. But in the campaigns during the last 
 twenty years for enlarging districts, states have offered certain 
 inducements for districts to unite. Such acts have been called 
 consolidation laws. The chief inducements offered are increased 
 powers and state aid. One of the new powers often conferred 
 upon such so-called consolidated districts is the authority to 
 transport pupils at public expense. 
 
 But in four states it is unlawful even in consolidated districts 
 to transport any pupils at public expense. In eleven other 
 states, boards of consolidated districts are authorized to trans- 
 port only those pupils living beyond a specified distance, two or 
 two and a half miles. And although eighteen states require 
 transportation of some or all children in consolidated districts, 
 the tax to sup])ly the funds must, in most states, be voted by 
 the people of the districts concerned. The courts^ have usually 
 held that a school board cannot be compelled, in such cases, to 
 transport children if the electorate has not voted the funds, and 
 transportation has not been as extensively provided as the state 
 laws indicate. State aid has not usually been given for trans- 
 portation, but it would seem, in view of the importance of the 
 distance factor in attendance, a desirable field for state partici- 
 pation. 
 
 Although improved roads may facilitate transportation, they 
 do not greatly reduce the distance for children who walk, a 
 mile of macadam road being equivalent to three-fourths of 
 a mile of dirt road. This advantage, however, is sufficient to 
 justify placing a new building on the main highway when the 
 building is not moved farther from the center of the popula- 
 tion to be served. The improved road makes the use of a motor-^ 
 
 ^ See Waters vs. State, 172 Ind. 251, 88 N. B. 67. 
 
 ' The laws of at least one state prohibit the use of a motor-bus in transporting 
 pupils. Note the details prescribed in the following, taken from Sec. 152, p. 33 of 
 the 1917 school laws of Oklahoma: 
 
 " • • . and in addition it shall be the duty of said district board to pro- 
 
 vide transportation to and from school for all pupils residing two or more miles 
 therefrom, in suitable vehicles of ample size, with comfortable seats, arranged to 
 conform to the size of the pupils to be carried, with adjustable covers for the com- 
 fort and protection of the pupils, drawn by stout, gentle teams, driven bv competent 
 persons of good moral character, who shall have control of the pupils during their 
 transportation." 
 
Adinmistrative Policies 25 
 
 bus in transportation practicable. Where transportation has 
 been tried out with different vehicles, experience has justified 
 the motor-bus where the roads are in good condition. 
 
 Since distance from school is not so important in the attend- 
 ance of older children, pupils in the higher elementary grades 
 might come together from a larger area, with separate one-room 
 schools for the younger children nearer their homes, if the popu- 
 lation is dense enough to permit it. But the distance factor is 
 important enough, even with the older children, to justify a 
 central school for all pupils if such school makes possible any 
 means of eliminating the distance factor. 
 
CHAPTER IV 
 THE MATERIAL AND METHOD USED 
 
 The Scope of the Data 
 
 An analysis of the factors influencing the attendance of country children 
 requires a knowledge of such facts of attendance as will enable one to 
 determine the kind of school attendance that accompanies the presence of 
 each factor. Since the problem cannot be tried out experimentally, it is 
 necessary to select as many factors as possible and to gather the facts of 
 attendance for them from a sufficient number of children to determine the 
 connection they have with attendance. Accordingly, the facts of attend- 
 ance in connection with facts about the children themselves, their schools, 
 and their communities, were gathered for the 6450 children enrolled in 
 the one-teacher country schools of five counties of Maryland. The blank 
 reproduced on the following page was used. 
 
 Facts about the Child 
 
 The facts about the child include his age, sex, grade in which enrolled, 
 the teacher's estimate of school work, whether promoted or retained at 
 the end of the year, how many years he has been in his present school, 
 how far he lives from the school, the kind of road he travels to school, 
 his attendance by months with the total for the year, the number of days 
 he walked, and the means of transportation used when he did not walk. 
 
 Facts about the School 
 
 The facts about the school include the training, salary, and experience 
 of the teacher and her rating for success in teaching as made by her 
 supervisor and county superintendent ; the number of recitations in the 
 daily program of the school; the games the children play; the meetings 
 of patrons held; the amount of money raised by the school for any pur- 
 pose; any special devices employed to secure regular attendance; the num- 
 ber of pupils enrolled and the number in average daily attendance; the 
 number of days the school was open by months, and the number of days 
 it was closed for various purposes; the size of the school building, the 
 number and location of windows, the kind of heating plant and its location, 
 the amount and kind of blackboard, the kind of desks and the amount of 
 teaching equipment, the size of the school grounds and the part of the 
 grounds used as a play ground. The county superintendent rated the 
 
 26 
 
Maryland One-Teacher Schools 
 
 Principars Annual Report 
 
 REPORT FOR THE SCHOOL YEAR 1917-18 
 Total number different pupils enrolled during the year? Boys Girls TotaL. 
 
 these who were not previously enrolled this year in any other school in this County' Boys- 
 
 Total number of 
 
 , Girls . Total 
 
 Number whu were not previously enrolled this year in any other school m this County or State? Boys Girls. Total 
 
 Total number enrolled over 16 years of age' Boys 
 Boys , Girls — 
 
 ., Girls ■■ . Total Total aggregate days' attendance' 
 
 Total Average daily attendance for the year' Boys Girls , Total 
 
 How many days was the school actually open for the instruction ol pupils during each calendar month ' September Octo- 
 ber November _.., December January.— — ., February March April May , 
 
 June , Total How many days was the school closed for teachers' meetings? How many days closed by order 
 
 of Board of Health' How many days closed by other governmental authority than the County Board of Education? 
 
 How jnany days was the school closed on account of sickness of teacher? How many days was it closed for other reasons' 
 
 For what reasons ?- — . ' 
 
 What are the dimensions of the classroom ' W idth in feet _, length What direction (north, south, east or west ) 
 
 do the children face when seated? ^ How many windows on the left of the children? on the right' _ How 
 
 many windows admit light in from of the pupils' How many from rear of pupils? Do the windows have curtains or 
 
 blinds? How many feet (length) of blackboard suitable (oi use? Of what material is the blackboard made, slate, 
 
 composition, wood or plaster? ._„ How many inches is the bottom of the blackboard from the floor? 
 
 What are ihe dimensions of the school ground, length in feet? . width in feet?.. 
 
 school ground is used by the pupils for a playground ' 
 
 About what fractional part of the 
 
 -If the pupils play outdoor games, name, in the order of their popu- 
 
 larity, the three games they most frequently play 
 
 Does the school have a good globe? How many wall maps suitable for use 
 
 In addition to the regular texts, how 
 
 many books sre there m the school library suitable for the children to use' — How many sets of supplementary readers for first- 
 grade pupils' For second-grade pupils? For third-grade pupils' 
 
 How IS your schoolroom heated, Smith system heater, Waterbury system heater, jacketed stove, unjacketed stove? 
 
 In what part of the room is the heater or stove? Is your schoolroom fitted with single or double 
 
 About how many years old are the desks? What pictures have you on the walls?. 
 
 How many recitations did you have each day this year when the attendance was greatest?.. 
 
 .Were there any meetings 
 
 of the patrons at the school house during the year' How many' --.. Did the school raise money for any purpose during 
 
 the year' _ How much' „_For what purpose'....... H you have found it necessary or advisable to use 
 
 any device or special method to encourage regular atendance this ytar, please describe such method or device 
 
 How many years have you taught ?- 
 
 How 
 
 Did you go to summer school in 1917'.. — It so wncre •'-.._ 
 
 many years in this school ' What is your annual salary including any bonus and any mcrease allowed during the present year? 
 
 Do you intend to continue teaching next year"'. _ 
 
 If you were employed in the State in December 1916 the State Superintendent has a record of your school training If you have 
 come into the service since that time please give an account of your schooling in the blanks provided below 
 
 Kind of 
 School 
 
 Name of School 
 
 1-ocation 
 
 of School 
 
 No Full 
 Years At-' 
 tendance. 
 
 P.irts nf 
 Years in 
 Months 
 
 Summer 
 Ses.«ions 
 in Weeks 
 
 Dhtes of Attendance. 
 
 Yenr of 
 Graduation 
 
 Klpuientary 
 School 
 
 
 
 
 
 
 
 
 High 
 School 
 
 
 
 
 1 1 
 
 
 Normal or 
 
 Training 
 
 Sclinol 
 
 
 
 
 
 
 
 
 Collpge or 
 University 
 
 
 
 
 
 
 
 
 Do not write below This space is reserved for the County Superintendent. 
 Teacher——^ . Building Grounds— _ Equipment. 
 
Page 2 
 
 LIST HERE ALL PUPILS WHO HAVE BEEN ENROLLEHJ AT ANY TLME DURING THE YEAR AND GIVE COMPLETE Jl 
 
 NAMES OF PUPILS. 
 
 9*^ 
 
 Date of 
 Entrance. 
 
 Date of 
 Withdrawal. 
 
 Promotf( 
 Retalr< 
 
3N EACH. BEFORE FILIJNG OUT THESE TWO PAGES, READ CAREFULLY THE DIRECTIONS ON PAGE FOUR. 
 
 Page 3-* 
 
 [liles from School 
 over different 
 kinds of roads. 
 
 DAYS ATTENDED DURING EACH CALENDAR MONTH. 
 
 S3 
 
 .MEANS OF TRANSPORTATION. 
 (Wlicn pupil did not walk.) 
 
LIST ON THIS PAGE ALL PUPILS FROM 6 TO If, VKARS OLD INCLUSIVE. AND LIYING IN THE SCHOOL DISTRICT. WHO 
 HAVE NOT BEEN ENROLLED THIB YEAR BEFORE FILLING OUT THIS PAGE, READ THE DIRECTIONS BELOW 
 
 
 3 
 
 a 
 
 o 
 CO 
 
 a 
 m 
 
 bo 
 
 < 
 
 When Last in 
 School. 
 
 a O 
 
 s ■S 
 
 £5 
 
 Miles from School 
 over dl£foreiit 
 kindd o( roads. 
 
 
 NAMES OF PUPILS. 
 
 > 
 
 a 
 
 u 
 a 
 
 a 
 
 a 
 D 
 
 
 CHIEF CAUSE OF ABSENCE 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ■ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 County 
 
 School No_ 
 Principal _ 
 
 Principal's Home Address. 
 
 Date of Making this Report 
 
 DIRECTIONS 
 In addition to the iniormation needed by the County Superintendent, this blank calls for some additional data that will be used by 
 the State Superintendent in an intensive study of school attendance in five representative counties This county has been selected by the 
 Sute Superintendent and it is hoped that the teachers will co-operate by carefully rcportinR all item* 
 
 1 Please answer all questions, leaving no blanks unfilled. 
 
 2 On page two, under "miles from school over different kinds of road" indicate the distance to the nearest 'A mile that the pupil 
 travels over improved ahd unimproved roads in going one way from home to school By "improved" road is meant macadam, con- 
 crete, shell, specially constructed sand-clay road or any other kind of specially constructed State or County road By "unimproved" 
 road is meant any ordinary road which has not been specially constructed by the County or State 
 
 3 B« careful to report "days attended during each calendar month" and the sum of these under "total days for the year" cor- 
 rectly and fully, as a part of the State school money is apportioned on this item 
 
 4 Estimate for each pupil the "number of days walked to school " It is not necessary to have this item exact Under "means of 
 transportation" indicate by a word or phrase the method of conveyance most commonly used when the pupil did not walk. That is. 
 did the pupil dnve or come on horseback, ride on auto bus, was he brought by father, brother, sister, or neighbor, or come on train 
 or trolley, and so on Please indicate the "means of transportation" even if the pupil was transported only two or three days. 
 
 5. In ichools where the teacher took the census, the names and ages for page four can be taken from the census report cards. 
 Note that in this report all ages are given as of September 1. 1917 while the census cards report ages as of September 1, I9ia 
 
 6 The "aggregate days attendance" is the suifl of the total days .ittendcci by each pupil The "average daily attendance" is found 
 ky dividing the aggregate days attendance by the number of days the school was in session. 
 
The Material and Method Used 27 
 
 buildings on their condition and fitness for school work, marking the best 
 fourth of all of them one, the second best fourth two, the next best 
 fourth three, and the final, or poorest fourth, four. He did this for 
 grounds, and equipment, also. 
 
 Facts about the Community 
 
 This study is not directly concerned with the economic or social level of 
 the community, as it attempts to determine the relation of modifications 
 of the school to the attendance of the children ; but to avoid the error 
 of drawing conclusions as to the connection of the teacher with attend- 
 ance, when the difference in the attendance might be due to differences 
 between the communities in which the teachers are located, it is necessary 
 to have some measure of the differences between communities in order to 
 isolate the community factor. It, therefore, seemed advisable to get an 
 estimate of the wealth, general intelligence, and permanent educational 
 interest of each neighborhood group. 
 
 The following excerpt from the letter of directions to the county super- 
 intendents for estimating the communities will explain how the rating was 
 made: 
 
 Divide the schools into four groups on the basis of the wealth, general 
 intelligence and permanent educational interest of the patrons. Number 
 the best group No. 1, the next group No. 2, the next No. 3, and the last 
 or poorest group No. 4. 
 
 There are, no doubt, places or parts of your county in which several 
 people cannot keep their children in school all year on account of the 
 need of clothing; they may be needed to work; or lack of interest in 
 education may cause some parents to keep their chidren out of school. 
 Think your county over and divide the schools into four groups on the 
 basis of these points, lumping the points together so as to think of them 
 as educational attitude and ability to keep children in school. 
 
 By permanent educational interest I mean a fixed interest. Some 
 teacher may stir up a breeze of enthusiasm in a community that does not 
 usually take much interest in education. Try to exclude this burst of 
 interest and think only of the deep-seated attitude of the school patrons. 
 
 In making this rating, think only of the parents of the children who 
 attend school. Disregard as far as you can the kind of teacher, kind of 
 building, equipment, etc., and consider only the school patrons. 
 
 In addition to the information called for on the blank, the offices of the 
 state and county superintendents were able to supply the expenditures by 
 schools, classified according to purpose, the total value of school property, 
 and which schools had changes of teachers during the year. 
 
 The Counties Selected 
 
 The blank on which this information was gathered was used as the 
 official annual report in the one-teacher schools of Allegany, Baltimore, 
 Calvert, Queen Anne's, and Wicomico Counties at the close of the school 
 year 1917-1918. Returns were received from all the one-teacher country 
 schools. These counties were selected because they include a wide range 
 of different conditions. 
 
28 Factors Controlling Rural School Attendance 
 
 Allegany County, in the western part of the state, is mountainous, with 
 some level country along the Potomac "bottoms" and along a stream in 
 the western part of the county. The soil is chiefly a clay loam. The lead- 
 ing industries are mining and manufacturing, except in the tillable sec- 
 tions of the county where farming is the chief industry. The county has 
 five high schools and a state normal school. It pays higher salaries and 
 has been able to secure better trained teachers than the state as a whole. 
 An assistant county superintendent is specifically charged with the super- 
 vision of the rural schools. The county has few state roads. 
 
 Baltimore County, extending from Baltimore City to the Pennsylvania 
 border, is the richest, and has the largest population of all the counties of 
 the state. The wealth of the county is not centered in the farming com- 
 munities, but the farmers are well-to-do. Several state roads lead into 
 Baltimore City, but almost half of the one-teacher schools are in districts 
 that have no state roads. The soil is such that an unimproved road be- 
 comes muddy in a rainy season. The county pays good salaries and has 
 paid one-room country school teachers a little more than the regular 
 schedule to make the positions in the country as inviting as those in the 
 suburban schools. For several years the board has employed only normal 
 school graduates, and has developed a system of supervision of instruction 
 that has attracted wide attention. A well trained assistant superintendent 
 devotes his full time to the rural schools. 
 
 Calvert County, on the western side of the Bay in the southern part of 
 the state, is the smallest county in the state, and one of the poorest finan- 
 cially. It is an agricultural region, with fishing and oystering along 
 the Bay. Except two schools which have a total of five teachers, all the 
 schools have only one teacher. The county has no high school. There is 
 one state road running north and south, the long way of the county, and 
 it has no railroad, except a branch line across the northern end of the 
 county. Until November, 1916, the county had a part-time superintendent 
 and practically no supervision. The present superintendent devoted his 
 first year chiefly to systematizing the organization and business side 
 of the schools, so that up to the time of this report the teachers had had 
 less professional direction than those in the other counties. The soil, 
 except in a few places, is a clay loam that produces mud in the unim- 
 proved roads during a rainy season. 
 
 Queen Anne's County has a fertile soil which has not been seriously de- 
 pleted by its 250 years of cultivation. Most of its rural schools are in 
 prosperous farming communities. Fishing and oystering are the chief 
 business of Kent Island and the western shore of the county. The county 
 has stressed good school buildings and equipment, and has teachers 
 of average training who have had superior supervision. The county has 
 several state roads, and a larger proportion of the country children than 
 in the other counties have some means of transportation. 
 
 Wicomico County has a light sandy soil, except in a section next to the 
 Delaware line which has a clay loam. The farming interests arc trucking 
 
The Material and Method Used 29 
 
 and fruit growing. Some of the rural school buildings, recently erected, 
 are of superior type. The teachers are of average training, are not so 
 well paid, but have had good supervision. 
 
 The school attendance law, uniform in all five counties, requires children 
 7-12 years of age to attend the full term. Thirteen and 14-year-old chil- 
 dren are required to attend 100 days, and the full term if they are not 
 lawfully employed. Children 15 and 16 years old, who have not com- 
 pleted the elementary school, must attend 100 days, and the full term if 
 not lawfully employed. Since it is unlawful to employ children under 14 
 years of age, 13-year-old children are required by the law to attend the 
 full term. 
 
 Each county has one attendance officer whose time is divided among 
 all the schools of the county. The efficiency with which the attendance 
 law is enforced is somewhat uniform.^ Baltimore County has employed 
 an attendance officer longer than the other counties and has placed more 
 stress upon the enforcement of the law. 
 
 Within each county the course of study is uniform. Baltimore County 
 has a well planned course of study that has received the favorable atten- 
 tion of school people over the United States. In the other counties the 
 course of study is worked out in less detail, but is on the whole up to 
 the general standard of rural schools. 
 
 Elimination of Transients 
 
 The names of the pupils who did not attend any days in May and June 
 were compared with the school census taken late in May to determine 
 whether they had moved from the school district. Late entrants were 
 checked by the column of the teacher's report which gave the number of 
 years each child had been in the school. In this way children moving in 
 or out of the district during the school year were located and dropped 
 from the lists. The children in the study include only those who lived in 
 the school district and were possible attendants during the entire year. A 
 total of 1135 names, representing about 700 different pupils, were dropped 
 from the lists, leaving 5315 children in the study. The proportion of 
 pupils moving, in the different counties, did not vary greatly. 
 
 A few of the items were so similar in the various reports, or so inde- 
 finitely reported, that no attempt v>'as made to treat them. For example, if 
 globes were furnished to some schools in a county, all scliools had them. 
 Under the head of special devices used to encourage attendance, few 
 teachers made any report and the statements of those reporting did not 
 lend themselves readily to quantitative treatment. The enrollment included 
 practically all children in the district wathin the prescribed ages, and no 
 analysis was made of the few, reported on the fourth page of the blank, 
 as not enrolled during the year. 
 
 ^ See Annual Report, Maryland State Board of Education, 1917, p. 32. 
 
30 Factors Controlling Rural School Attendance 
 
 The Preliminary Survey of the Data 
 
 The volume of the data and the labor involved in its analysis made it 
 advisable to drop any factor from further consideration as soon as it was 
 known to have no fixed relation to attendance. Accordingly, a preliminary 
 survey of the data was made in order to eliminate certain possible factors, 
 and to select others for further study. 
 
 The distribution of the attendance, though skewed towards the higher 
 end, was spread enough at the ends to allow two samples, one taken at 
 either end of the distribution, to be selected, which showed a marked dif- 
 ference in attendance. By taking a sample from the top of the distribution 
 in each county to get the children who attended best, and another sample 
 from the bottom to get those who attended poorest, and comparing the two 
 samples thus selected, it should be found that any fact that varies with 
 attendance to any appreciable extent should appear in dififerent amounts in 
 the two samples. 
 
 Although the state law prescribes a minimum school term of 180 days, 
 the schools in the five counties varied considerably in the number of days 
 they were open during the year. To avoid the error of selecting an 
 undue proportion from the schools open a larger number of days, the 
 top sample was selected by taking those attending a given percentage of the 
 days their schools were open. The bottom sample was selected by taking 
 all attending less than a given percentage of the days their schools were 
 open. The samples taken from the top of the distribution are called "G." 
 The samples from the bottom of the distribution are called "P." The G 
 children are those who attend most ; the P children are those who attend 
 least. 
 
 For example, out of 1231 children in Wicomico County, 167 were 
 present 90 per cent of the days their schools were open, and 176 children 
 attended only 45 per cent, or less, of the days their schools were open. 
 Any factor that varies with attendance, such as the efficiency of the 
 teacher, should be found in different amounts in these two groups of 
 children. If the better teacher as a rule has better attendance, a larger 
 percentage of those children attending best should be found under the 
 better teachers; and, if living farther from the school has a tendency 
 to decrease attendance, more of those children who attended least should 
 be found living farther from the school. Two samples, each including 
 about 12 per cent of all in the county, were selected from each county 
 without any effort to get the samples equal in numbers, but care was 
 taken to draw a straight line through the distribution including within the 
 sample all that were cut off by the line. The two samples were then tabu- 
 lated according to a wide range of possible factors. The facts that showed 
 no connection with attendance were discarded, and those with a significant 
 relation to attendance were selected for further study. For example, 
 Table 1, reproduced below, shows how the two groups of pupils were 
 distributed according to the handwriting of the teachers in three counties. 
 
The Material and Method Used 31 
 
 The teacher with the superior handwriting seems to have a very slight, if 
 any, advantage over the teacher with inferior handwriting in the distribu- 
 tion of the two groups of children. In county number 2, more of the 
 best-attending (G) pupils and fewer of the poorest-attending (P) 
 pupils, are under the teachers with superior handwriting. There is, how- 
 ever, no difference in the other two counties. A factor that has a fixed 
 and significant relation to attendance should show a more constant ten- 
 (lency. 
 
 TABLE 1 
 
 Showing the Distribution of Both Attendance Groups 
 
 According to the Handwriting of the Teacher 
 
 Handwriting of Teacher 
 
 Superior Average Inferior Total 
 
 CY 1 G 52 105 32 189 . 
 
 P 56 98 34 188 
 
 CY 2 G 58 99 46 203 
 
 P 43 92 67 202 
 
 CY 5 G 32 92 43 167 
 
 P 33 98 45 176 
 
 Total G 142 296 121 559 
 
 P 132 288 146 566 
 
 Table 2, on the following page, shows the quality of work of the two 
 groups of pupils. It is clear that attendance and quality of work are 
 closely related. In the light of the facts of Table 1 the teacher's hand- 
 writing was discarded as a factor in attendance, and in view of the 
 facts of Table 2 the relation of quality of work to attendance was selected 
 for further study. 
 
 The Analysis of Significant Factors 
 
 The facts that showed an important relation to attendance were made 
 comparable by assigning scores to such facts as were not already ex- 
 pressed in quantitative terms. The relation between the several factors 
 was found by tabulating the children with reference to the several pairs 
 of factors and computing the coefficients of correlation by the method of 
 unlike-signed pairs.^ 
 
 The attendance and distance curves were considerably skewed and the 
 application of a highly refined method would have been a difficult task on 
 
 1 Thorndike, E. L., Mental and Social Measurements, Teachers College, Columbia 
 University, 1913, p. 170. 
 
32 Factors Controlling Rural School Attendance 
 
 account of the volume of the data. In choosing between covering a 
 narrower range of facts with a more refined measure and a wider field 
 with a simple formula of approximation, the wider field was chosen. The 
 important factors were tabulated for three age groups in each of the five 
 counties, giving fifteen coefficients between each pair of factors. The 
 coefficients did not vary greatly, and in each case distinctly showed a 
 central tendency. 
 
 TABLE 2 
 
 Showing the Distribution of Both Attendance Groups 
 
 AcCOltDlNG TO THE QuAUTY OF ScHOOt WoRK 
 
 Quality of School Work 
 
 Superior Average Inferior Total 
 
 Cy 1 G 44 108 30 182 
 
 P 8 60 95 163 
 
 Cy 2 G 48 140 15 203 
 
 P 12 59 109 180 
 
 Cy 3 G 34 50 6 90 
 
 P 7 60 69 136 
 
 Cy 4 G 8 53 5 66 
 
 P 2 70 34 106 
 
 Cy 5 G 42 104 12 158 
 
 P 5 90 66 161 
 
 Total G 176 455 68 699 
 
 P 34 339 Z7Z 746 
 
 Witli so many variables, the factors were so interrelated that it was 
 impossible to determine the connection between any two of them by the 
 ordinary coefficient of correlation. They were isolated by partial cor- 
 relation. The need for partial correlation can be shown by an illustration. 
 
 For example, Holley found that the education of parents, the number of 
 books in the home, and the amount of rent paid, were directly related to 
 the amount of schooling the children in a family receive; but the parents 
 with more education usually have larger home libraries, and more often pay 
 a higher rental ; and his material does not show which of these factors is 
 responsible for their connection with the education of the children ; or 
 whether parents with more education, more books, and better homes did 
 not come from a different original stock, and being from such stock, 
 more often have the kind of children who stay longer in school and reach 
 higher grades. 
 
 Van Denburg found that the children of Russian Jews of New York 
 City go twice as far in high school as the children of Irish descent, with 
 fifteen times larger proportion of the Jews remaining to complete 
 
rhe Material and Method Used 33 
 
 the course. A study of clothing merchants as compared with policemen 
 might show that the children of clothing merchants have greater per- 
 sistence in high school than the children of policemen, when the occu- 
 pation of the parents had nothing to do with high school persistence, the 
 difference being due to the fact that clothing merchants are more often 
 of Jewish extraction while policemen are more often Irish. 
 
 To get a correct estimate of the influence of the racial factor, per sc, 
 it would be necessary to compare similar samples of the two races. We 
 should select samples from the two races that possess the same amount 
 of all other important factors. The two race groups should have similar 
 incomes, similar living expenses, families of similar size, similar edu- 
 cational opportunities in their homes, and be alike in any other way 
 that tends to keep their children in or out of school, except in the factor 
 of race. 
 
 It would, however, be very diflScult, if not impossible, to find two sam- 
 ples with a simple factor isolated. But two such samples may be com- 
 pared with respect to a single factor, free from its connection with other 
 factors, by partial correlation. Any factor can be eliminated or held 
 constant by partial correlation. 
 
 The formula ^ is : 
 
 f — r r 
 
 12 13 '23 
 
 "■'~ 1/ (l-r^3) {l-r\_,) 
 in which r^.,.., indicates the correlation between factors 1 and 2 for a 
 constant value of factor 3. By the continued application of the formula 
 one factor afte/ another may be eliminated or made constant, so that the 
 relation between any two variables, freed from their connection with any 
 others, may be determined. The labor involved and the degree of accu- 
 racy required has prevented the application of this method to a group of 
 several variables; but the use of Kelley's " tables and a mechanical cal- 
 culating machine greatly reduces the labor, making the task comparatively 
 easy. 
 
 Students of education have not usually given in their published re- 
 ports the cross correlations necessary for the application of the formula 
 of partial correlation. In a few instances Holley gives sufficient data. 
 
 Applying the formula to the following coefficients of correlation in his 
 study, 
 
 Books in Education of 
 the home fathers 
 
 Education of sons 67 .47 
 
 Books in the home .60 
 
 1 Yule, G. Undy, An Introduction to the Theory of Statistics, London, 1916, 
 Chap. 12. 
 
 ^Bulletin of University of Texas, No. 27, 1916. 
 
34 Factors Controlling, Rural School Attendance 
 
 we find the correlation between the education of sons and the education 
 of fathers, for all sons and fathers who have the same number of books in 
 their homes, drops from .47 to .11; while the correlation between the edu- 
 cation of sons and books in the home, for fathers with any fixed or given 
 amount of education, drops only from .67 to .56. This would indicate that 
 the .47 correlation between education of sons and education of fathers 
 was largely due to the fact that each of these two factors is tied up with 
 the factor of books in the home. That is, educated fathers do not more 
 often have educated sons so much because the fathers are better educated 
 as because both better educated sons and better educated fathers are 
 more often found in the homes of the sort of people who have more 
 books. 
 
 Again applying the formula to 
 
 Education of Education of 
 fathers mothf.rs 
 
 Education of sons 47 .55 
 
 Education of fathers .65 
 
 the correlation between education of fathers and education of sons, consid- 
 ering the mothers' education constant, drops from .47 to .18; while the cor- 
 relation between the education of sons and education of mothers, keeping 
 the education of the fathers constant, is .36. This shows that part of 
 the correlation between the education of sons and fathers is due to the 
 connection of the education of each with the education of the mother. 
 That is, educated fathers more often have educated sons because they 
 more often have educated wives, who in turn more often have educated 
 sons. Of course, a part of the correlation between the education of 
 mothers and the education of sons is accounted for by the slight connection 
 between the education of the fathers and sons. From Holley's data it is 
 possible to make a similar treatment of books in the home, the education 
 of daughters, the education of fathers, and the education of mothers. The 
 education of mothers seems to have a considerably stronger connection 
 with the education of the children than the education of the fathers has. 
 and it is probable, if the economic and other factors were eliminated, the 
 slight remaining connection between education of fathers and children 
 would largely disappear. And the factor of the mother's education, already 
 shown to be less important than the factor of books in the home, would 
 be less after it is freed from its connection with other factors. Had 
 Dr. Holley isolated the several factors he studied, it is very probable he 
 would not have inferred that "the amount of education of the parents is 
 the most important and persistent factor influencing the schooling of the 
 children." ^ 
 
 1 Hollev, C. E., The Fifteenth Year Book, National Society for the Study of 
 Education, Part VI, p. 103. 
 
TJic Material and Method Used 35 
 
 Summary 
 
 1. The facts of attendance with the desired facts about the children 
 themselves, their schools, and an estimate of their communities were 
 gathered for the 6450 children enrolled in the one-teacher schools of 
 five counties of Maryland. 
 
 2. The counties were selected so as to get samples of a wide range of 
 different conditions. 
 
 3. The form on which this information was gathered was made the official 
 annual report for the year 1917-1918 and was filled out in full by all 
 the teachers. 
 
 4. The data were supplemented by data from the offices of the state and 
 county superintendents. 
 
 5. The samples of the best-attending and poorest-attending children from 
 each county were tabulated for a wide range of possible factors, and 
 the significant factors selected for further analysis. 
 
 6. Coefficients of correlation were computed for the several pairs of sig- 
 nificant factors and the independent relation of the factors determined 
 by partial correlation. 
 
CHAPTER V 
 
 PRl'LIAIINARY SURVEY OF THE DATA 
 
 In the preliminary treatment of the data, the factors influencing the 
 school attendance of the children will be considered under three heads: 
 the child, the school, and the community, followed by a discussion and 
 summary. As explained in Chapter IV, samples of the best-attending 
 and poorest-attending children were taken from the top and bottom of the 
 distribution in each county, and the two groups compared with respect to 
 the several factors.^ The two samples are indicated by the letters G and P 
 respectively. 
 
 The Child and Attendance 
 Sex as a Factor 
 
 The G (best-attending) groups of children were about evenly divided 
 between boys and girls in four counties, with the girls exceeding 
 the boys in the fifth county ; but there were more than twice as many 
 boys as girls in the P (poorest-attending) groups. While it is not 
 shown in any table here, it was apparent in making the tabluations that this 
 difference among the P children is chiefly due to better attendance by 
 older girls than by older boys. 
 
 Age as a Factor 
 
 The median of all the G children was little more than three years 
 younger than the median of the P children. This difference varied from 
 two to four years in the five counties. A larger proportion of the G 
 groups are from eight to eleven years old. Few of the G children are 
 thirteen or over, and a larger percentage of the P children are twelve and 
 over. Children 6-7, 8-11, and 12 and over seem to form the most natural 
 attendance age-groups. (Table 3) 
 
 The Grade in Which Enrolled 
 
 Although the median of all the G children is three years younger than 
 the median P child, they are practically together in their school work. 
 The median G pupil is ahead in two counties and behind in three. In the 
 counties in which the G pupil is behind, he is also younger, so that the 
 median P child remains, for his age, about three grades behind the G child. 
 (Table 4) 
 
 ^ I'or lack of space only a few of the tabulations made in the preliminary survey 
 of the data are given. A complete set of the tables are on file in Bryson Library, 
 Teachers College, Columbia I'niversity, New York City. 
 
 36 
 
Preliminary Survey of the Data 37 
 
 Quality of Work 
 
 Each group of children was classified according to the quality of their 
 school work as reported by the teacher. Those ranking in the top fifth 
 of all pupils were classed as "superior," the middle three fifths "average," 
 and the bottom fifth "inferior" in quality of work. Ten per cent of the 
 G children were inferior in school work as compared with 50 per cent of 
 the P group. Twenty-five per cent of G children were superior as com- 
 pared with 5 per cent of the P children. A marked relation between at- 
 tendance and quality of work is found in all the counties. (Table 5) 
 
 The Relation of Promotion to Attendance 
 
 Ninety per cent of the G children were promoted as compared with 
 25 per cent of the P children. Most of the failures among the G children 
 came from a few schools in which an unusual per cent of the children was 
 ''retained." For example, the seven G children in Calvert County who 
 failed were in one school (School 9, District 2) which promoted only 12 
 out of 35 children. (Table 6) 
 
 TABLE 3 
 
 Showing the Distribution of the Two Groups of Children 
 
 According to Age 
 
 CY 1, CY 2. CY 3, CY 4, and CY 5 are used to designate the counties, 
 
 Allegany, Baltimore, Calvert, Queen Anne's, and Wicomico, respectively. 
 
 G is used to designate the group of children who attended best. P is used 
 
 to designate the group of children who attended poorest. 
 
 Age of Pupils 
 
 6 7 8 9 10 11 12 13 14 IS Total 
 
 CY 1 G 17 24 22 31 2,i 29 14 9 2 1 182 
 
 P 15 14 11 12 12 10 19 19 16 35 162 
 
 CY 2 G 30 27 27 28 23 29 23 9 5 2 202 
 
 P 13 6 10 12 8 4 14 36 35 42 180 
 
 CY 3 G 3 4 6 18 14 12 8 12 8 5 90 
 
 P 12 13 8 5 5 5 10 16 28 28 130 
 
 CY 4 G 14 17 23 17 14 10 8 5 2 3 113 
 
 P 9 6 7 1 6 13 10 19 24 17 112 
 
 CY 5 G 16 17 31 25 21 20 19 7 4 .. 160 
 
 P 14 12 12 9 13 13 22 27 25 29 176 
 
 Total G 80 89 109 119 105 100 72 42 21 11 748 
 
 P 63 51 48 39 44 45 75 117 128 151 761 
 
38 Factors Controlling Rural School Attendance 
 
 TABLE 4 
 
 Showing the Distribution of the Two Groups of Children 
 According to the Grades in Which They WpjtE Enrolled 
 
 Grades 
 
 12 3 4 5 6 7 8 Total 
 
 CY 1 G 30 22 40 27 29 24 9 1 182 
 
 P 59 21 23 10 24 10 12 4 163 
 
 CY 2 G 31 34 33 36 26 25 11 7 203 
 
 P 34 13 21 20 23 37 15 17 180 
 
 CY 3 G 10 6 13 15 17 16 13 2 92 
 
 P 41 6 13 21 26 14 15 1 137 
 
 CY 4 G 20 17 19 24 15 10 8 .. 113 
 
 P 23 11 12 16 21 11 22 .. 116 
 
 CY 5 G 14 24 32 27 30 19 20 . . 166 
 
 P 33 14 21 30 26 28 24 .. 176 
 
 Total G 105 103 137 129 117 94 61 10 756 
 
 P 190 65 90 97 120 100 88 22 772 
 
 TABLE 5 
 
 Showing the Distribution of the Two Groups of Children 
 According to the Quality of Their School Work 
 
 Quality of School Work 
 
 Superior Average Inferior Total 
 
 CY 1 G 44 108 30 182 
 
 P 8 60 95 163 
 
 CY 2 G 48 140 15 203 
 
 P 12 59 109 180 
 
 CY 3 G 34 50 6 90 
 
 P 7 60 69 136 
 
 CY 4 G 8 53 5 66 
 
 P 2 70 34 106 
 
 CY 5 G 42 104 12 158 
 
 P 5 90 66 161 
 
 Total G 176 455 68 699 
 
 P 3-* 339 373 746 
 
Preliminary Suwey of the Data 39 
 
 TABLE 6 
 
 Showing the Number of Children in Each Group Who Were 
 Promoted at the End of the Year 
 
 Promoted 
 
 Yes No Total 
 
 CY 1 G 164 13 177 
 
 P 23 107 130 
 
 CY 2 G 173 17 190 
 
 P 29 125 154 
 
 CY 3 G 37 7 44 
 
 P 27 50 77 
 
 CY 4 G 96 12 108 
 
 P 27 50 77 
 
 CY 5 G 144 21 165 
 
 P 23 105 128 
 
 Total G ; 614 70 684 
 
 P 129 437 566 
 
 TABLE 7 
 
 Showing the Distribution of Both Attendance Groups According to 
 THE Distance the Pupils Live from the School 
 
 Distance from the School in Miles 
 
 —% —'A —1 —lyi —2 —3+ Total 
 
 CY 1 G 121 27 24 5 4 1 . . 182 
 
 P 24 14 37 20 34 30 4 163 
 
 CY 2 G 84 40 52 16 5 6 . . 203 
 
 P 16 14 50 34 51 15 .. 180 
 
 CY 3 G 48 13 16 4 6 5 . . 92 
 
 P 17 12 17 21 2>7 29 4 137 
 
 CY 4 G 32 25 18 9 14 12 3 113 
 
 P 8 9 25 15 17 26 13 113 
 
 CY 5 G 66 19 44 24 7 5 2 167 
 
 P 15 11 SO 61 25 12 .. 174 
 
 Total G 351 124 154 58 36 29 5 757 
 
 P 80 60 179 151 164 112 21 767 
 
40 Factors Controlling Rural School Attendance 
 
 TABLE 8 
 
 Showing the Distribution of the Two Groups of Children According 
 
 TO THE Training of the Teachers in Years Above the 
 
 Elementary School 
 
 Training in Years 
 Above the Elementary School 
 
 12 3 4 5 6 7 Total 
 
 CY 1 G 4 16 46 43 22 58 . . 189 
 
 P 20 44 9 39 11 65 .. 188 
 
 CY 2 G 13 .. 7 21 91 30 38 3 203 
 
 P 6 9 20 107 41 12 7 202 
 
 CY 3 G 7 19 12 8 30 5 1 10 92 
 
 P 14 12 36 15 47 7 3 3 137 
 
 CY 4 G 3 27 59 19 5 . . 113 
 
 P 7 34 33 25 12 5 116 
 
 CY 5 G 39 21 69 . . . . 129 
 
 P 13 16 51 68 .. .. 148 
 
 Total G 20 23 38 141 244 145 102 13 726 
 
 P 20 32 109 94 277 152 92 15 791 
 
 TABLE 9 
 Showing the Distribution of the Two Groups of Children According 
 
 TO THE x\M0UNT OF PROFESSIONAL TRAINING OF ThEIR TeACHERS 
 
 Professional 
 Training in Years 
 
 ^ 14 1 - Total 
 
 CY 1 G 26 31 8 37 87 189 
 
 P 55 14 4 15 100 188 
 
 CY 2 G 22 10 .. 12 159 203 
 
 P 30 22 .. 2 148 202 
 
 CY 3 G 9 36 16 7 24 92 
 
 P 25 24 26 15 47 137 
 
 CY 4 G 36 39 29 9 113 
 
 P 37 51 3 1 24 116 
 
 CY 5 G 8 28 22 60 11 129 
 
 p 13 42 22 66 5 148 
 
 Total G 101 144 75 116 290 726 
 
 P 160 153 55 99 324 791 
 
CY 
 
 1 
 
 G 
 P 
 
 CY 
 
 2 
 
 G 
 P 
 
 CY 
 
 3 
 
 G 
 P 
 
 CY 
 
 4 
 
 G 
 P 
 
 CY 
 
 5 
 
 G 
 P 
 
 Total 
 
 G 
 
 
 
 P 
 
 Preliminary Survey of the Data 41 
 
 TABLE 10 
 
 Showing the Distribution of the Two Groups of Childrex 
 
 ACCORDING to the SALARIES PaID ThEIR TeACHERS 
 
 Salary of the Teacher 
 
 $350 400 450 500 550 600 650 700 750 800 850 Total 
 
 20 8 17 72 23 .. 28 21 189 
 
 41 19 37 52 20 11 8 188 
 
 29 42 32 IS 20 25 24 203 
 50 38 45 10 8 13 7 202 
 
 2 
 
 
 6 
 
 8 
 
 29 
 
 3 
 
 11 
 
 11 
 
 6 
 
 50 
 
 
 56 
 
 2 
 
 9 
 
 12 
 
 
 83 
 
 1 
 
 36 
 
 7 
 
 42 
 
 11 
 
 33 
 
 5 
 
 16 
 
 10 
 
 33 
 
 39 
 
 17 
 
 1 
 
 28 
 
 101 
 
 25 
 
 
 10 
 
 48 
 
 112 
 
 2 
 
 5 
 
 3 
 
 79 
 127 
 
 107 
 100 
 
 164 
 170 
 
 92 176 83 94 90 42 60 36 20 25 24 742 
 
 102 258 90 116 81 49 53 10 8 13 7 787 
 
 TABLE 11 
 
 Showing the Distribution of the Two Groips of Children 
 
 According to the Rating of Their Teachers 
 
 IQ means the best fourth of the teachers; 2Q the next best fourth, etc. 
 
 Rating of Teachers 
 lO 20 30 40 Total 
 
 CY 1 G 36 69 37 47 189 
 
 P 48 43 41 56 188 
 
 CY 2 G 
 
 P 
 
 CY 3 G 
 P 
 
 CY 4 G 
 P 
 
 CY 5 G 
 P 
 
 Total G 
 P 
 
 69 
 
 60 
 
 46 
 
 28 
 
 203 
 
 53 
 
 58 
 
 48 
 
 43 
 
 202 
 
 21 
 
 8 
 
 21 
 
 42 
 
 92 
 
 41 
 
 10 
 
 25 
 
 61 
 
 137 
 
 33 
 
 44 
 
 35 
 
 1 
 
 113 
 
 34 
 
 18 
 
 28 
 
 36 
 
 116 
 
 60 
 
 47 
 
 35 
 
 25 
 
 167 
 
 34 
 
 45 
 
 50 
 
 47 
 
 176 
 
 219 
 
 228 
 
 174 
 
 143 
 
 764 
 
 210 
 
 174 
 
 192 
 
 243 
 
 819 
 
4.2 Factors Controlling Rural School Attendance 
 
 The School and Attendance 
 Training of the Teacher 
 
 The two groups of children were tabulated (Table 8) according to the 
 years of school training above the elementary school wliicli their teachers 
 had. A slightly larger proportion of the G children was found under the 
 better trained teachers in the several counties, but the teacher of the 
 median of all G children had only about one-tenth of a year more schooling 
 than did the teacher of the median P child. There seems to be only a 
 slight connection between tlie total schooling of the teacher and the attend- 
 ance of her pupils. 
 
 The reports of the teachers in regard to their professional training were 
 verified and supplemented by the records of the State Department of Ed- 
 ucation, which keeps a cumulative card index of each teacher's scholastic 
 and professional preparation. 
 
 The term "professional training," as here used, includes not only courses 
 in educational psychology, principles of teaching, methods, school organi- 
 zation, and the like, lying strictly within the field of education, but also 
 normal school courses in subject matter, when the emphasis was upon 
 how to teach the subject. The teachers with more professional training 
 have about the same proportion of each group of the children. Teachers 
 without training have a larger proportion of the P children, but the median 
 of all G children is under a teacher with slightly less professional training 
 than the median P child. (Table 9) 
 
 Under the Maryland law, teachers' certificates are of three grades : The 
 first grade certificate is granted to applicants who have completed a stand- 
 ard two years' normal school course in addition to a four years' high 
 school course. A number of teachers holding first grade certificates, how- 
 ever, completed the normal school course when the requirements did not 
 include four years of high school work for entrance. A second grade 
 certificate is issued to four-year high school graduates. A third grade 
 certificate is issued to applicants with less than a high school education. 
 Second and third grade certificates require six weeks' professional train- 
 ing in a teacher-training institution. Temporary certificates of the different 
 grades are issued to a sufficient number of applicants to fill the schools after 
 the supply of regularly certified applicants is exhausted. 
 
 Some teachers have life certificates issued under a former law. The 
 life certificates do not represent a definite amount of school training, but 
 usually indicates a longer and satisfactory service. 
 
 The grade of certificate seems to have no direct bearing upon attend- 
 ance, except that the life certificate holders have a larger proportion uf the 
 G children. This may be due, in part, to the preferred communities in 
 which their longer service has placed them. 
 
Preliminary Survey of the Date 43 
 
 Salary of the Teacher 
 
 The salary schedule is uniform for each county and is determined by the 
 grade of certificate, the number of years of experience, and the rating for 
 success in teaching given by the county superintendent. In all the counties 
 the teachers who were paid higher salaries had a slightly larger proportion 
 of the G children, and a few more of the P children were under the teach- 
 ers receiving less. The median of all G children and 450 of the P children 
 were under teachers receiving less than $500, while 207 of the G group 
 and 140 of the P group were under teachers receiving $600 or more. This 
 difference may be due, in part, to the longer tenure of the higher paid 
 teachers, many of whom are in the preferred communities. (Table 10) 
 
 Experience of the Teacher 
 
 In four of the five counties, beginning teachers had a smaller proportion 
 of the G children than of the P children, with a total of 212 G, as compared 
 with 305 P children, under teachers with one year or less of experience. 
 Beginning with the third year of experience, there is little difference ex- 
 cept in Allegany and Baltimore Counties, where the teachers of long-time 
 service have a larger proportion of the G children. 
 
 This information was gathered from the school year 1917-18 when teach- 
 ers were leaving for more lucrative positions with the government in Wash- 
 ington, and for positions in industries in nearby cities. The counties 
 could not compete with the government and industry in the inducements 
 offered, and the professional level of incoming teachers was, no doubt, 
 lower than had been maintained in the past. On this account, beginning 
 teachers, as a class, could hardly be compared with those longer in the 
 service. Again, those longer in the service had the advantage in their 
 choice of communities, so that the difference between the number of G 
 and P children found under beginning and old teachers could not be as- 
 cribed to experience alone. 
 
 Number of Recitations 
 
 The teacher's organization of the day's work, as expressed in the num- 
 ber of recitations in the daily program of the school, has a small relation 
 to the distribution of the two groups of children. The schools with fewer 
 recitations have a slight advantage in three counties. On the whole, schools 
 with 27 recitations or less have more of the G children, and schools with 
 28 recitations or more have more of the P children. This difference is 
 more marked in schools having 40 recitations or more. 
 
 Completeness of the Teacher's Report 
 
 A few of the teachers failed to report all of the facts asked and it 
 was necessary to write them for the missing items. The reports of these 
 teachers were marked "incomplete" and the two groups of children tabu- 
 
44 Factors Controlling Rural School Attendance 
 
 lated with reference to whether they were found under teachers returning 
 complete or incomplete reports. There seems to he no relation between 
 the attendance of the pupils and the completeness of their teacher's report. 
 
 Neatness of the Teacher's Report 
 
 The reports varied consideralily in neatness and were divided into 
 three groups on this basis. The reports were rated by three judges and 
 the median estimate taken as the final score for tlie neatness of each re- 
 port. There seems to be no relation between attendance and the neatness 
 of the teacher's report. 
 
 Handwriting of the Teacher 
 
 The teacher's reports from three counties were scored according to the 
 legibility of their writing. In Baltimore County the teachers whose writ- 
 ing w-as superior had a slightly larger proportion of the G children, but 
 there was no difference in the other two counties. The quality of a teach- 
 er's handwriting seems to have no relation to the attendance of her pupils. 
 
 Teacher'.s Intention to Continue Teaching 
 
 A slightly larger portion of the G children were under teachers who 
 expressed an intention to continue teaching. While this difference is small, 
 it is found in four of the five counties, there being practically no difference 
 in Baltimore County. 
 
 Schools Changing Teachers 
 
 On account of the increased demands for help in the government service 
 and industrial pursuits, a larger number of teachers than usual gave up 
 their positions in the schools during the year. The two groups of children 
 were tabulated in four counties according to whether they were in schools 
 that changed teachers during the year. In all four counties, the schools 
 changing teachers had a few less of the G children and a few more of the 
 P. There was a total of 71 G children as compared with 126 P children in 
 schools changing teachers, with 601 G children and 556 P children in the 
 schools that did not change teachers during the year. 
 
 Rating of the Teacher 
 
 In four of the five counties the better teachers, as rated by the county 
 superintendent and the supervisor for success in teaching, had a larger 
 proportion of the G children. In Calvert County the teacher rated highest 
 had a slightly smaller proportion of the G children. The better half of the 
 teachers in all counties had 447 G and 384 P children while the poorer half 
 of the teachers had a total of 317 G and 435 P children. (Table 11) 
 
Preliminary Survey of the Data 45 
 
 School Library 
 
 The schools were divided into three classes according to the size of the 
 libraries. The best fourth was classed as superior, the middle half as 
 average, and the poorest fourth as inferior. In three counties the schools 
 with superior libraries had a larger proportion of the G children, and 
 schools with inferior libraries had a larger proportion of the P children. 
 There was not so wide a range of difference between the libraries in the two 
 other counties, each school having a fair library. The size of the school 
 library seems to have a small, but not marked relation to school attendance. 
 
 Blackboard 
 
 The kind of blackboard, the amount of blackboard, the amount of black- 
 board per pupil attending five months, and the height of the bottom of the 
 blackboard from the floor apparently have little, if any effect upon attend- 
 ance. A school with less than sixteen linear feet of blackboard, with less 
 than four tenths of a linear foot per pupil, or with the blackboard within 
 24 inches or less from the floor, may have a slightly smaller proportion of 
 the G pupils. But the blackboard may be considered a negligible factor. 
 
 Rating of the Equipment 
 
 The superintendent of each county divided the schools into four equal 
 groups on the basis of the amount and quality of the teaching equipment. 
 In two counties the half of the schools with better equipment had a larger 
 proportion of the G cliildren, in two counties there was no difference, and 
 in one county the schools with better equipment had fewer G children. 
 The amount and kind of teaching equipment do not seem to be significant 
 factors. 
 
 Age of Desks 
 
 In two counties the schools with the older desks had a slightly larger 
 proportion of the G children, in two others they had a smaller proportion, 
 and in one county there was no difference. On the whole, the age of the 
 school desks does not seem to be a factor in school attendance. 
 
 Pictures on the Walls 
 
 The schools in each county were divided into two groups on the basis of 
 the number and kind of pictures on the walls. The quality of the pictures 
 was passed upon by three experienced teachers of extended training, and 
 the median mark taken for each school. In two counties the schools hav- 
 ing more and better pictures had more of the G children, in two others 
 they had more of the P children and in the fifth county there was no 
 difference. In the total for all the counties, the schools with better pic- 
 tures had a slightly larger proportion of the G children. 
 
46 Factors Controlling Rural School Attendance 
 
 Size of the School 
 
 In four of the five counties the two groups of children were in schools 
 of practically the same enrollment. In Allegany County more of the G 
 children were in the schools with larger enrollment. The median of all G 
 children was in a school enrolling 34 pupils, and the median P child in a 
 school enrolling Z2i. 
 
 Taking the number of pupils on roll who attended at least part of five 
 different months as a measure of the size of the school, there is no differ- 
 ence in two counties, but in the other three more of the G children were in 
 the larger schools. The median of all G children was in a school of 9, 
 and the median P child in a school of 28. The school having more of the 
 G pupils would have more attending five different months, as the better 
 attendance itself would tend to increase the number in school five months. 
 
 In all five counties the median children were in schools having higher 
 average daily attendance, but the differences were not great. From a com- 
 parison of all the G children with all the P children, it is clear that a larger 
 proportion of the G children is found in the schools having higher averag'. 
 daily attendance. The median of all G children was in a school having an 
 average daily attendance of 22; the median P child was in a school with 
 an average attendance of 19. The schools with more G pupils would tend 
 to have a higher average attendance as a result of these star pupils. From 
 the slight relation between the number enrolled in the various schools and 
 the distribution of the two groups of children, the comparatively small 
 increase in the average daily attendance and the number attending five 
 months in the schools with more of the G children, the size of the school 
 may be considered, at most, only a slight factor. 
 
 Size of the Schoolroom 
 
 From the dimensions of the schoolroom the floor space was computed 
 and the two groups of children tabulated according to the size of the 
 schoolrooms in which they were located. The larger buildings in three 
 counties had a larger proportion of the G children; there was no difference 
 in the other two counties. The median of all G children was in a room 
 having 625 square feet, and the median P child was in a room with 601 
 square feet of floor space. 
 
 The floor space was divided by the number of children in school five 
 months, and the two groups of children tabulated according to the floor 
 space per pupil. In three counties a larger proportion of the G children 
 were in schools having a larger amount of floor space per pupil, and in 
 two counties there was no difference. The median of all G children was 
 in a school that had 22 square feet per pupil, and the median of all P 
 children in a school with 22 square feet. 
 
Preliminary Survey of the Data 47 
 
 Lighting of the Schoolroom 
 
 More than one half of each group of children is in school buildings with 
 light admitted from two sides; two per cent of the entire number is in 
 buildings with light on four sides. The buildings with light on one and 
 three sides more often had a larger proportion of the P children, but 
 these differences are not very marked. 
 
 Many buildings erected more than ten years ago have one or more win- 
 dows in front of the pupils. A model, widely used, has one window in the 
 centre of the room directly behind the teacher's desk. On the whole, a 
 school without windows in front of the pupils has a better chance to get a 
 larger proportion of the G children, but the difference is insignificant. 
 
 The direction in which the seats face seems to have practically nothing 
 to do with attendance. So far as there is a difference, rooms facing east 
 and south have a larger proportion of the G children than tlie rooms facing 
 north or west. 
 
 Heating and Ventilation 
 
 The schools were classified according to the system of heating and venti- 
 lation used. Those with modern systems, such as the Smith or Waterbury, 
 which heat and ventilate as well, were placed in one group ; those with 
 jacketed stoves in another; and schools with a plain stove in another. The 
 heating system seems to have little, if any relation to the attendance. If 
 there is any difference, the schools with the modern systems have a larger, 
 and the schools with jacketed stoves have a smaller proportion of the G 
 children. The schools were again classified according to the location of 
 the heater. The place of the heater in the room seems to have no connec- 
 tion with the attendance in four of the five counties. In Queen Anne's 
 County the schools with the heaters in the end of the room have a larger 
 proportion of the G children. 
 
 Rating of the School Building 
 The county superintendent divided the schools into four equal groups 
 according to the condition and fitness of the school buildings. In two 
 counties the schools with better buildings had a larger proportion of the G 
 children; in two counties there was no difference; and in the fifth county 
 the better buildings had a smaller proportion of the G children. 
 
 School Grounds 
 The county superintendent divided the schools into four equal groups 
 according to the suitability of the school grounds. In two counties a 
 larger proportion of the G children was in the schools having better 
 grounds ; in two counties a larger proportion was in the schools with in- 
 ferior grounds; and in the fifth there was no difference. In the total for 
 all counties the children of the two groups were about evenly distributed, 
 so that the kind of grounds is not a considerable factor in school attend- 
 ance as measured by the distribution of these two groups of children. 
 
48 Factors Controlling Rural School Attendance 
 
 Play and Games 
 
 The schools were divided into two approximately equal groups on the 
 basis of the play and games reported. One group of schools played more 
 games, and plaj'ed more together; the other group of schools reported 
 fewer games, and the games were more individual. The play of each 
 school was rated by three judges, and the median judgment taken. In 
 four of the five counties, the schools marked higher for play and games 
 had a larger proportion of the G children. The total of all P children was 
 about equally divided, but 433 of 764 G children were in the schools that 
 had more play and games. 
 
 Value of School Propekty 
 
 The office of the county board of education keeps a fairly accurate and 
 complete inventory of all school property, and the two groups of children 
 in three counties were tabulated according to the total value of the prop- 
 erty of the schools in which they were located. The median G child in 
 all three counties was in a school with more valuable property. The dif- 
 ference was small in two counties ; in the third the median G child was 
 in a school with property worth about 8 per cent more than the property of 
 the school containing the median P child. 
 
 Cost of Running the School 
 
 The accounting system used by the county boards of education is uni- 
 form for the state. Costs are kept by schools according to the purpose of 
 the expenditures, under the classification recommended by the N. E. A. 
 and the United States Bureau of Education. The two groups of children 
 were tabulated according to the current or running expense, which in- 
 cludes the total expenditure for each school less the amount paid on debt 
 service and for capital outlay. The two largest items in the running ex- 
 pense of a one-teacher school are the salary of the teacher and the cost 
 of fuel. In three counties the median G child was in a school with a 
 higher running expense than the median P child. In two counties there 
 was no difference. In Baltimore County, which shows the greatest differ- 
 ence, the school of the median G child cost about 5 per cent more than the 
 school of the P child. The median of all G children was in a school cost- 
 ing $588, and the median P child in a school costing $539. 
 
 The current expense less the teacher's salary, the chief items of which 
 are fuel, janitor's wages, and supplies of various kinds, including text 
 books, was tabulated for the two counties in which current expense bore 
 the closest relation to attendance. There is no relation between expendi- 
 tures for these purposes and the distribution of the G and P children in the 
 schools of either county. 
 
Preliminary Siinrey at the Data 49 
 
 Distance from the Children's Homes 
 
 60 per cent of the G children (Table 7) live within ^ mile of the 
 school, while 80 per cent of the P children live more than 14 ™ile from the 
 school. The median of all G children lives approximately a quarter of a 
 mile from the school, while the median P child lives approximately a mile 
 and a quarter from the school. The action of the distance factor is very 
 significant and is similar in all counties. 
 
 The Community and Attendance 
 
 Money Raised 
 
 The funds for supporting the schools in Maryland are furnished jointly 
 by the county and the state. The local school district does not tax itself 
 for school purposes, but the law provides that the county board of educa- 
 tion must furnish an equal amount when a school raises a given sum for 
 additional library books. Schools often raise money to purchase musical 
 instruments or other equipment not supplied to all schools by the county. 
 The two groups of children in each county were tabulated according to 
 the amount of money raised. In three counties the schools raising more 
 money had a considerably larger proportion of the G children. There 
 was no apparent difference in the other two counties, but the schools in 
 these two counties raised less money than those in the other three counties. 
 
 Community Meetings 
 
 The schools were divided into two groups on the basis of the number 
 of community meetings held during the year. Schools in Baltimore County 
 holding one meeting or less, and schools in Wicomico County holding two 
 meetings or less, were classed as holding "fewer meetings." "Fewer meet- 
 ings" in the three other counties means no meetings at all;, but some 
 schools in all counties held from one to twenty meetings. In three of the 
 counties, a slightly larger proportion of the G children was in the schools 
 which held more meetings; in one county there was no difference; and 
 in one county the schools holding more meetings had less of the G children. 
 
 In the total for all counties, the schools with more meetings had a 
 larger proportion of the G children. 
 
 Rating of the Community 
 
 The superintendent of each county divided the schools into four equal 
 groups on the basis of the wealth, general intelligence, and permanent 
 educational interest of the patrons. The schools in the better communities 
 of three counties had a considerably larger proportion of the G pupils, 
 there being no significant difference in the other two counties. The three 
 counties — Allegany, Baltimore and Wicomico — in which the community as 
 rated by the county superintendent has some relation to attendance, are 
 
50 Factors Controlling Rural School Attendance 
 
 also the counties in which the amount of money raised by the schools 
 showed a similar relation. In the total. 227 G and 221 P children were m 
 the bottom fourth of all the communities. 
 
 Discussion and Summary 
 The Child 
 
 Boys and girls in country schools are similar in their attendance except 
 that boys 12-14 years old attend a smaller number of days than girls of 
 the same ages. The attendance from 8 to 11 is fairly constant. Children 
 6-7 attend a smaller number of days than children 12 and over. Six-year- 
 old children do not go as many days as those seven years old, and there 
 is a gradual decline in each year beginning at 12 years. The 12 per cent of 
 all children with poorest attendance are three years older than the 12 per 
 cent of all children who attend best but are in the same grade. The chil- 
 dren attending best are rated considerably higher in quality of work, and 
 are much more often promoted. 
 
 Children 5-7, 8-11, and 12 form the most natural age groups; attend- 
 ance seems to have a definite relation to the grade in which a child is 
 enrolled for his age, and the quality of his school work. These factors 
 will be treated further in the next chapter. 
 
 The School 
 
 The training, salary, and rating of the teacher seem to have some con- 
 nection with attendance, and will receive further consideration. After the 
 first year, the experience of the teacher apparently has no relation to at- 
 tendance, except in two counties where teachers of long-time service have 
 more of the children who attend best. Since the difference in the attend- 
 ance under the beginning teachers may be due to a lower quality of new 
 teachers on account of the war-time demands for labor, and the slightly 
 better attendance under the long-tenure teachers may be due in part to 
 the preferred communities in whicli their long service has placed them, 
 the experience of the teacher does not have an important relation to 
 attendance. As the length of service enters into the amount of salary 
 paid, the experience of the teacher will be included, so far as its importance 
 seems to justify, in a treatment of the teachers' salaries. 
 
 The number of recitations, the completeness and neatness of the teach- 
 er's report, and the legibility of her handwriting have apparently no re- 
 lation to attendance. There is some evidence that the teachers not in- 
 tending to continue in the service have poorer attendance, and the attend- 
 ance in the schools changing teachers during the year is even a little poorer. 
 The dissatisfied teacher leaving the service was usually replaced with a 
 teacher with less training and experience; but the training and experience 
 of the teacher are not important factors in attendance, and may be even 
 less responsible for the low attendance than the kind of communities in 
 which the changes occur. 
 
Preliminary Survey of the Data 51 
 
 The county superintendent's rating of the equipment, the size of the 
 hbrary, and the play and games of the children seems to have a small rela- 
 tion to attendance; but the amount and kind of blackboard, the age of 
 the school desks, the size of the school, the size of the school room, the 
 lighting, the heating system, the pictures on the walls, and the rating of 
 the school grounds have no significant relation to school attendance. 
 
 The fact that there is less connection between attendance and the total 
 running expense than between attendance and the teacher's salary indi- 
 cates that fuel, repairs and other things included in the remainder of 
 running expense are less important factors in attendance than is the 
 teacher. The current expense, less the teacher's salary, in the schools of 
 the two counties which showed the closest relation between current expense 
 and attendance, had no relation to attendance when treated by itself. 
 
 The kind of teacher, as determined by the salary paid, training, and 
 rating for success in teaching, will be considered further. The physical 
 provision for the child, including building, equipment, and grounds, seems 
 to bear little, if any relation to attendance ; but, on account of the im- 
 portance of these items in our scheme of education, they will be further 
 analyzed. 
 
 The Community 
 
 The kind of community in which a school is placed, as measured by 
 the community meetings, the money raised, and estimated by the county 
 superintendent, seems to have some relation to the attendance of the 
 children. The community meetings held and the money raised are per- 
 haps as much a result of the teacher's activity as of the intelligence and 
 educational interest of the community. The county superintendent is in a 
 good position to estimate the permanent educational interest and attitude 
 of the school patrons, and his estimate of the community is perhaps more 
 reliable than any other measure available. Tn the remainder of the study 
 it will be used as a measure of the difference between communities. 
 
CHAPTER VI 
 ANALYSIS OF SIGNIFICANT FACTORS 
 
 The preliminary survey of the data (Chapter V) showed that the dis- 
 tance children live from the school, the grades in which they are enrolled 
 in relation to their ages, and the quality of work they do have a definite 
 relation to school attendance. The teacher, the school plant, and the 
 educational interest of the school patrons also seemed worthy of further 
 consideration. Having selected certain factors concerning the child, the 
 school, and the community for further study, the significant facts in re- 
 gard to each must be made comparable. 
 
 Scoring thk Child 
 
 It is necessary to express the age-grade relation of each child by a quan- 
 titative measure, and also to translate the teachers' reports on quality of 
 work into comparable quantitative units before these factors can be re- 
 lated to the other significant facts known about the children and their 
 schools. 
 
 A child who fails of promotion twice in four years, and is two grades 
 behind after four years in school, has not progressed so well as one 
 who has failed only twice in six years, and who is two grades behind at 
 the end of six years in school. Although the first child has progressed at 
 the better rate than one who has failed the same number of times in 
 three years, the latter will perhaps fail again during the next three years 
 and be three or four grades behind at the end of six years in school. Like- 
 wise, it means more to be promoted every year for six years than to 
 "escape failure" for a shorter period. 
 
 To do "satisfactory work" in the proper grade for one's age means 
 more than to do the same quality of work in a grade behind one's 
 age; but it means less than to do the same quality of work in a grade ahead 
 of one's age. The age-grade relation and quality of work are tied up 
 together and should be combined in some way in order to get the best 
 estimate of the achievement of each child. 
 
 Accordingly, scales were devised to weight the age-grade relation and 
 quality of work of each child so that the two facts might be combined into 
 a single measure of achievement. In any combination of the two facts 
 for an index of achievement, the mark for quality of work should per- 
 haps receive more weight in the lower grades, and the age-grade relation 
 
 52 
 
Analysis of Significant Factors ' 53 
 
 more emphasis in the higher grades. The relative weight for each fact 
 could be determined statistically ; but for lack of time this relation 
 was not computed, and in the absence of definite information the two 
 factors were given equal weight in making a single score for achieve- 
 ment. Each was assigned values ranging from to 10, with an average 
 of about 5 for each factor. 
 
 Queen Anne's and ^Vicomico Counties have seven grades in their coun- 
 try schools, Allegany and Baltimore Counties have eight, and Calvert 
 County has a primary grade followed by eight grades, making nine. The 
 counties vary also in the amount of retardation among the children at 
 different ages, so that a scale for scoring the agp-grade relation in one 
 county would not apply to the others. An age-grade table was tabulated 
 for each county and weights assigned to children of each age according 
 to the distribution of the children of that age througliout the different 
 grades. 
 
 The score for any age-grade group was determined roughly by the per- 
 centage of children of the given age between the group and the median of 
 all the children of that age. The age-grade table and the scale for scor- 
 ing the age-grade relation of the children in Baltimore County will serve 
 to illustrate the method used (Tables 12, 13). For example, the median 
 11 -year-old child in Baltimore County was in about the middle of the fifth 
 grade, and the 48 11-year-old children in the fifth grade were assigned an 
 age-grade score of 5. The Z7 children of the same age in the fourth grade 
 were assigned a score of 4; the 12 in the third grade, a score of 2; the 
 7 in the second grade, a score of 1 ; arid the three in the first grade, a score 
 of 0; while the 36 11-year-old children in the sixth, the 10 in the seventh, 
 and the 1 in the eighth grade were assigned scores of 7, 10 and 12, re- 
 spectively. The scores for age-grade relation of all the 11-year-old chil- 
 dren in Baltimore County average 5.1. 
 
54 
 
 Factors Controlling Rural School Attendance 
 
 u 
 
 o 
 
 % 
 
 H 
 
 < 
 
 
 o 
 
 13 
 
 w ;i; 
 
 -r r^ 00 rvi <^ "T? '^'2 >o 
 
 -M "-* ^ (N ^ •-' ^ 
 
 -H'^JrOrfu-jsOr^OO 
 
 CM 00 "^ -^ On 
 
 ^ • lo r^ "^ lo IT) 
 
 ;0 r^ t^ 
 •^ ii-> -—> 
 
 ro "^ lO 
 
 o 
 
 « 
 in 
 
 rt '^l CO 
 
 t^ >J^ -H 
 
 00 ■* 
 
 5 
 
 O iri 
 
 -^rMoo'^to^'^00 
 
.'.iialysis of Sigiiiiicaut Factors 55 
 
 TABLE 13 
 
 Showing the Scale for Weighting the Age-Grade Relation of 
 Baltimore County Pupils, as Developed from the Age-Grade Table 
 
 Grade in 
 
 
 
 
 
 
 Age 
 
 OF 
 
 Pupils 
 
 
 
 
 
 
 WHICH enrolled 
 
 5 
 
 6 
 
 7 
 
 8 
 
 9 
 
 10 
 
 11 
 
 12 
 
 13 
 
 14 
 
 15 
 
 16 
 
 17 
 
 18 
 
 1 
 
 6 
 
 5 
 
 4 
 
 3 
 
 2 
 
 1 
 
 
 
 
 
 
 
 
 
 2 
 
 
 9 
 
 6 
 
 4 
 
 3 
 
 2 
 
 1 
 
 
 
 
 
 
 
 
 3 
 
 
 
 9 
 
 6 
 
 5 
 
 1 
 
 2 
 
 1 
 
 1 
 
 
 
 
 
 
 4 
 
 
 
 
 9 
 
 7 
 
 5 
 
 4 
 
 2 
 
 2 
 
 1 
 
 
 
 
 
 5 
 
 
 
 
 
 9 
 
 7 
 
 5 
 
 4 
 
 3 
 
 2 
 
 1 
 
 
 
 
 6 
 
 
 
 
 
 
 10 
 
 7 
 
 6 
 
 5 
 
 3 
 
 2 
 
 2 
 
 
 
 7 
 
 
 
 
 
 
 12 
 
 10 
 
 8 
 
 7 
 
 5 
 
 4 
 
 3 
 
 2 
 
 
 8 
 
 
 
 
 
 
 
 12 
 
 10 
 
 8 
 
 7 
 
 6 
 
 5 
 
 3 
 
 2 
 
 This average is reduced gradually for older children and is increased 
 for children 5-7 years old. It means more to get an early start and less 
 not to be through the elementary school at the higher ages. It is probable 
 that the children entering school at an early age are more often the bright 
 children and more mature for their ages. It is also probable that more 
 of the older children still in the elementary school are dull, more of the 
 Iiright children having passed on to other schools. The early entrants 
 can be considered the top of the distribution and the remaining older 
 children the bottom of the distribution of all the children of those ages. 
 
 To eliminate the varying standards of different teachers in marking 
 their pupils, the quality of work was translated into scores in a similar 
 way. The mark of the median child in school was assigned a value of 5, 
 and the marks of the groups of other children scored from to 10 ac- 
 cording to the percentage of children between the group receiving i 
 given mark and the median child in their school. Most of the teachers 
 reported the quality of work by words such as excellent, superior, good, 
 fair, poor, and the like; but this method of scoring quality work is appli- 
 cable to any system of marks which discriminates between different quali- 
 ties of work. The assignment of scores in three representative schools of 
 Queen Anne's County illustrates the method used. (Table 14) 
 
 After the age-grade relation and quality of work of each child in the 
 five counties had been scored, the two scores for each child were added 
 to make a single score for achievement. 
 
 Scoring the School 
 
 It is also necessary to have a scale to score each school, if the school is 
 considered as a single factor. The country school can be treated as com- 
 posed of two elements: the teacher and the school plant. The preliminary 
 
50 Factors ControU'uii!; Rural School Attendance 
 
 TABLE 14 
 
 Show[ng the Scores Assigned for Quality of Work in Three 
 Representative Schools of Queen Anne's County 
 
 School District Mark used uy No. of children Scores assignf.d 
 
 THE teacher receiving THE MARK FOR THE MARK 
 
 10 1 Very Good 5 9 
 
 Good 19 5 
 
 Fair 8 2 
 
 Excellent 
 
 4 
 
 8 
 
 Very Good 
 
 7 
 
 6 
 
 Good 
 
 8 
 
 4 
 
 Fair 
 
 4 
 
 2 
 
 Excellent 
 
 1 
 
 10 
 
 Very Good 
 
 5 
 
 8 
 
 Good 
 
 10 
 
 6 
 
 Fair 
 
 9 
 
 2 
 
 Poor 
 
 1 
 
 
 
 survey of the data (Chapter V) shows that the teacher is a more impor- 
 tant factor in school attendance than the building, equipment, and 
 grounds. Tlicre are but two facts about the teacher that show any con- 
 siderable connection with attendance — her rating by the county superin- 
 tendent and her salary, — although her total school training has some rela- 
 tion to attendance, and the best mark may be a combination of the three 
 elements, giving each a weight according to tlie strength of its connec- 
 tion with attendance. The combination of the three elements which gives 
 the highest correlation with attendance places the correct value on each 
 element. 
 
 After further study of Tables 8 and 11, several coml)inati.)ns of these 
 elements were tried out and scores were assigned as follows : A teacher 
 was given 9, 6, 3, and 0, respectively, for being rated in the best fourth, 
 second fourth, third fourth, and poorest fourth of all the country teachers 
 in her county ; 1 point for each $50 received in salary above the legal 
 minimum of $350, and 1 point for each year of training above the elemen- 
 tary school. This weighting may give a little more emphasis to training 
 than the preliminary analysis seems to justify, but training is more constant 
 
 than the other two variables, and the three factors seem to combine in 
 these proportions to make the score that correlates highest with at- 
 tendance. 
 
 In a similar way a scale was developed for scoring the school plants. 
 The preliminary survey showed little if any relation between most of the 
 physical facts of the school and attendance, and after several attempts 
 
Analysis of Significant Factors 57 
 
 to get a scale that would include a wider range, it was finally limited to 
 the following, which, after trial, seemed to produce the most significant 
 mark for the school plant: A building was given 5, 3, 2, and 0, respec- 
 tively ; the grounds 3, 2, 1, and 0, respectively, for being ranked in the 
 best fourth, second fourth, third fourth, and poorest fourth of all its 
 class in the country schools of the county. The scores assigned on these 
 three points were added to get a single mark for each school plant. The 
 scores of the teachers vary from 1 to 22, with an average of about 12. 
 The scores for the school plants vary from to 10, with an average 
 slightly below 5. The two scores were then added to get a mark for each 
 school. 
 
 The children were tabulated by age-groups and counties in correlation 
 tables for attendance, with achievement, school, and distance, and the co- 
 efficients of correlation computed by the method of unlike-signed pairs. ' 
 The coefficients show a fairly high positive correlation between attendance 
 and achievement, an equally high negative correlation between attendance 
 and distance from school, and a low correlation between attendance and 
 school. (Table 15) 
 
 TABLE 15 
 
 Showing Attendance Correlated with Achievement of Pupils, the 
 Kind of School, and Distance from School, by Age- 
 Groups AND Counties 
 
 Attendance Correlated With 
 
 CouxTv Achievement School Distance 
 
 5 8 5 8 5 8 
 
 Ar.E-CiROUP to tr> to to to to 
 
 7 11 12-1- 7 11 12+ 7 11 12-1- 
 
 CY 1 54 .54 .50 .14 .30 .21 -.58 .54 -.17 
 
 CY 2 53 .44 .57 -.02 .19 -.06 -.39 -.36 -.30 
 
 CY 3 68 .51 .65 -.01 -.11 -.02 -.52 -.55 -.44 
 
 CY 4 41 .46 .40 .29 .18 .06 -.48 -.37 -.23 
 
 CY 5 49 .53 .59 .22 .18 .24 -.37 -.34 -.10 
 
 The Correction of Distance 
 
 An ;• of 1 indicated perfect correlation; that is, variation in which no 
 chance variation occurs. An r of indicates variation wholly by chance or 
 such a balance between contributing factors that the influences neutralizo 
 each other. Since errors and other variations due to chance tend to lower a co- 
 efficient of correlation, the comparatively high negative correlation between 
 attendance and distance would be even higher if errors and variations in the 
 teachers' judgment of distance were eliminated. Some of the children go 
 to school over macadam state roads, and others must travel unimproved 
 
 ^ Tliorndike, E. L., Mental and Social Measiiretnents, p. 170. 
 
58 J'actors ControUing Rural School Attendance 
 
 dirt roads tliat hccome muddy and impede travel durint? a rainy season. As 
 measured by tlie effort it tjikes to get to school, a child living a mile from 
 school on a state road is not so far away as one who lives the same dis- 
 tance on an unimproved road. Treating similar distances on improved and 
 unimproved roads as equal has the same effect on results as errors in 
 the judgment of distance. A more accurate relation between attendance 
 and distance from school would be shown if the distances over one kind 
 of roads were translated into proper amounts of the other kind of road. 
 
 Since more children go to school over unimproved roads, distances over 
 state roads were translated into corresponding amounts of dirt roads. In 
 Baltimore County there were 300 children 7-13 years old in 16 schools hav- 
 ing state roads in their school districts. Some of the children in each of 
 these schools went all or part of the way over state roads, and some went 
 a part or all of the way over dirt roads. 
 
 TABLE 16 
 
 SniiwiNG Median and Quartiles for Attendance and the Median Dis- 
 tance OF Pupils by Age-Groups and Counties After the Distances 
 Were Corrected for Kind of Roads and Days Transported 
 
 County Age-Group Attfxdance in Days Medi.vn 
 
 IQ M 3Q Distance in Milks 
 
 CY 1 5-7 75 117 147 .49 
 
 8-11 88 127 153 .47 
 
 12 plus 42 77 120 .75 
 
 Total 65 110 146 .56 
 
 CY 2 5-7 96 129 154 .68 
 
 8-11 104 132 154 .74 
 
 12 plus 43 87 134 .83 
 
 Total 79 121 149 .76 
 
 CY 3 5-7 72 108 137 .69 
 
 8-11 104 125 147 .91 
 
 12 plus 64 106 141 .85 
 
 Total 83 117 141 .84 
 
 CY 4 5-7 124 155 169 .82 
 
 8-11 126 156 172 .87 
 
 12 plus 69 125 159 .91 
 
 Total 109 149 168 .86 
 
 CY 5 5-7 102 127 140 .84 
 
 8-11 104 128 148 .80 
 
 12 plus 58 94 139 .81 
 
 Total 89 122 142 .81 
 
Analysis of Significant Factors 59 
 
 The coefficient of correlation between attendance and uncorrected dis- 
 tance for the 300 children was found to be .38. 
 
 Three guesses were made on the relation between the improved and un- 
 improved roads, the distances on state roads corrected according to each 
 guess, the attendance of the children tabulated by the new distances, and 
 the coefficients of correlation figured for each guess. Considering distance 
 over a state road equal to one-fourth the distance over a dirt road gave a 
 correlation of .38, the same as that obtained with no correction at all ; taking 
 the relation as one-half, the correlation was .43 ; and counting distance 
 over a state road equal to three-fourths as much distance over dirt roads 
 gave a correlation of .54. 
 
 A guess on either side of three-fourths was then tried. Five-eighths 
 gave .48, and seven-eights gave .51. The three guesses (5/8, 3/4, and 7/8) 
 were tried out on the other four counties by selecting the children that 
 would be displaced in the correlation tables by each correction and com- 
 puting the effect of the displacements on the coefficients of correlation. 
 Counting a state road as equal to three-fourths of as much dirt road was 
 the best correction in each county. 
 
 Likewise, a few children have some means of transportation which 
 tends to overcome the distance factor. A child living three miles from 
 school who has an opportunity to come regularly in a comfortable con- 
 veyance is not so far from school, as measured by the effort it takes to get 
 to school, as a child living the same distance who must walk. To count 
 them the same distance in considering the relation of the distance factor to 
 attendance would have the same effect upon the results as an error on the 
 part of the teacher in estimating distance. A more accurate relation will be 
 found if the distance of transported children is reduced according to the 
 advantage they have by reason of their means of transportation. The cor- 
 rection in the distance of transported children that gives the highest corre- 
 lation between attendance and distance, when the groups of transported 
 and non-transported children are treated together, is the correct relation 
 between the two. 
 
 In Baltimore County there were 371 children 6-12 years old in schools 
 with no state roads and in which one or more children were transported 
 5 per cent or more of the days attended. Several corrections for trans- 
 portation were tried out on these children, reducing the distance different 
 amounts for various per cents of the days attended that the children were 
 transported. The following seemed to express the relation best : no cor- 
 rection for transportation 5 per cent or less of the days attended ; five 
 sixths of the distance for transportation 6-10 per cent of the days at- 
 tended; four fifths of the distance for transportation 11-20 per cent of the 
 days attended ; three fourths of the distance for transportation 21-40 per 
 cent of the days attended; one half of the distance for transportation 
 41-70 per cent of the days attended ; and one third of the distance if trans- 
 ported 71-100 per cent of the days attended. When applied to the trans- 
 ported children, the correction raised the correlation between attendance 
 and distance for the 271 children from .40 to .55. 
 
60 Factors Controlling Rural School Attendance 
 
 TABLE 17 
 
 Showing the Correlation of Attendance and Achievement 
 With Both Corrected and Uncorrected Distance 
 
 ^ Achievement correlated Attendance correlated 
 cd c WITH Distance with Distance 
 
 ^ Q Uncorrected Corrected Uncorrected Corrected 
 
 f S-7 -.58 -.77 -.25 -28 
 
 CY 1 ] 8-11 -.54 -.60 -26 -27 
 
 [ 12+ -.19 -.21 -.12 -.17 
 
 r5-7 -.2,7 -A7 -.22 -.31 
 
 CY 2 j 8-1 1 -.36 -.45 -.25 -.27 
 
 [ 12+ -.25 -.32 -.25 -.31 
 
 r5-7 -.52 -.54 -.12 -.13 
 
 CY 3 8-11 -.37 -.44 -.16 -.16 
 
 I 12+ -.44 -.48 -.25 -.27 
 
 r5-7 -.48 -.61 -.23 -.25 
 
 CY 4 ] 8-11 -27 -.61 -.06 -.20 
 
 I 12+ -.23 -.43 -.09 -.13 
 
 I' 5-7 -.35 -.43 -.08 -.16 
 
 CY 5 j 8-11 -.39 -.42 -.29 -.32 
 
 [ 12+ -.11 -.19 -.03 -.09 
 
 This correction, however, underestimates the effect of transportation on 
 distance, because it gives an advantage to transported children who 
 attended a small number of days. For example, a child who attended only 
 20 days, with some means of transportation for ten of them, would get as 
 large a reduction in distance by the formula as one attending 180 days with 
 90 days of transportation; yet the distance and lack of transportation the 
 remainder of the time might have been the cause of the first child's long 
 absence. To reduce this advantage, the correction was reduced one half 
 for children attending less than 100 days, and again one half for children 
 attending less than 60 days. In applying the correction to the group 12 
 years of age and older, the median days of attendance was substituted for 
 the 100 days in the correction formula when the median days attended was 
 less than 100. 
 
 A correction, worked out on either the total number of days transporterl 
 or the percentage the transported days were of the days the school was 
 open, would give an undue advantage to transportation ; and the method 
 used was selected because its fault is conservation rather than exaggeration. 
 
Analysis of SigniUcant Factors Gl 
 
 TABLE 18 
 
 Showing the Crude or Original Coefficients of Correlation, 
 Using the Corrected Distances 
 
 Distance. Age Gr. Qual. of W. Teacher. Phys. Pro. Com. 
 
 5-7 12+ 5-T 8-11 8-11 12+ 7^-' 8-11 12+ o-7 8-11 12+ 5-7 8-11 12+ 5-7 8-11 12+ 
 
 y Cr 1 -.77 -.60 -.21 .41 .54 .01 .40 .35 .30 .01 .13 .17 .16 .14 .00 .07 .34 .17 
 
 a CY 2 -.47 -.45 -.32 .38 .45 .02 .34 .33 .47 .08 .23 .19 -.03 .07 -.22 .31 .30 .37 
 
 "3 Cr 3 -.54 -.44 -.48 .37 .45 .51 .66 .44 .61 .08 .09 -.12 .06 .00 .00 .08 .08 .02 
 
 ;2 CY 4 -.61 -.61 -.43 .20 .50 .39 .41 .16 .30 .36 .27 .18 .05 -.17 -.03 -.13 -.13 .06 
 
 < CY 5 -.43 -.42 -.19 .59 .58 .55 .13 .33 .48 .21 .16 .22 .16 .08 .04 .24 .34 .45 
 
 -.19 -.31 -.20 -.13 -.08 -.09 -.17 -.21 -.41 .10 .09 .07 -.10 -.34 .-14 
 
 -.14 -.32 -.28 -.24 -.17 -.25 .00 -.10 .08 .00 .03 .07 -.03 -.18 -.17 
 
 -.33 -.19 -.24 -.13 -.13 -.18 .16 .15 .04 .09 -.13 .03 .11 .02 .04 
 
 -.11 -.20 -.10 -.17 -.06 -.18 -.03 -.03 .07 .09 -.06 -.06 .13 .06 .13 
 
 -.17 -.19 -.11 -.12 -.15 .04 .00 -.13 -.04 -.19 -.13 -.08 -.02 .05 .09 
 
 .25 .30 .42 -.18 -.13 .02 .03 .09 -.04 .25 .20 .24 
 
 .02 .23 .66 -.09 .01 .05 .02 .17 -.08 .13 .05 .20 
 
 .41 .30 .54 -.02 -.03 -.03 -.09 -.02 .17 .04 .22 .00 
 
 .15 .24 .53 .24 .10 .13 .11 .08 -.03 -.04 .01 .23 
 
 .29 .11 .32 .02 .02 .04 .06 -.07 .00 .25 .08 .19 
 
 .06 -.09 -.10 .11 .08 -.09 .07 .07 -.03 
 .00 .22 .08 -.08 .12 .02 -.03 .09 .17 
 -.18 .05 -.08 -.23 .02 .07 -.14 .03 -.05 
 .18 -.06 .00 -.15 -.16 .08 -.10 -.17 .03 
 .14 .00 -.04 .10 .11 .06 -.10 -.03 .00 
 
 .00 .00 .00 .33 .33 .33 
 
 .o2 .32 .32 .34 .34 .34 
 
 .25 .25 .25 .37 .37 .37 
 
 .00 .00 .00 .42 .42 .42 
 
 .39 .39 .39 .35 .35 .35 
 
 .00 .00 .00 
 
 .27 .27 .27 
 
 .12 .12 .12 
 
 .17 .17 .17 
 
 CY 5 . .79 .79 .79 
 
 The corrections for kind of roads and days transported were made for 
 all children displaced by these corrections in the correlation tables of each 
 county, and the new coefficients of correlation computed. The correlations 
 between attendance and distance, and between achievement and distance 
 were raised in every case and in somewhat similar proportions (Table 17). 
 The corrected distance was used in all further tabulations. 
 
 The Crude Coefficients 
 
 The correlations between attendance, the score for age-grade relation, 
 and quality of work seemed to indicate that the age-grade relation and 
 
 
 CY 
 
 1 
 
 a 
 
 CY 
 
 2 
 
 5 
 
 CY 
 CY 
 CY 
 
 3 
 4 
 
 5 
 
 0^ 
 ■J 
 
 CY 
 CY 
 CY 
 
 1 
 3 
 
 
 CY 
 
 4 
 
 «« 
 
 CY 
 
 5 
 
 "u. 
 
 CY 
 
 1 
 
 ■? 
 
 CY 
 
 2 
 
 "S 
 
 CY 
 
 3 
 
 ■:; 
 
 CY 
 
 4 
 
 5 
 
 CY 
 CY 
 
 5 
 1 
 
 .3 
 
 CY 
 CY 
 
 2 
 3 
 
 
 CY 
 CY 
 
 4 
 5 
 
 6 
 
 CY 
 OY 
 
 1 
 2 
 
 1 
 
 CY 
 
 3 
 
 CY 
 
 4 
 
G2 Factors Controlling^ Rural School Attendance 
 
 quality of work are more independent of each other than was assumed 
 when they were combined into a single score for achievement. 
 
 The correlation between attendance, teacher, and school plant showed a 
 similar relation between the teacher and the school plant. Since these 
 factors are more significant when treated separately, they were treated as 
 factors in the remainder of the study. The correlations between the sev- 
 eral pairs of the remaining factors were computed, using the corrected 
 distance. These correlations, by age-groups and counties, are shown in 
 Table 18. 
 
 The correlation between attendance and distance is around 50, being 
 higher for the 5-7 age-group and decreasing with each of the older groups. 
 The decrease between the 5-7 and the 8-11 groups is slight, but the de- 
 crease between the 8-11 and 12 groups is more marked. 
 
 The correlation between attendance and age-grade relation is just about 
 equal to the correlation between attendance and distance, but is of opposite 
 sign, being positive. It also varies with age, but gets larger with each 
 successive group, the largest jump being between the 5-7 and 8-11 groups. 
 There is about .38 for 5-7, and .50 for 8-11, and .55 for the 12 groups. 
 
 The r's between attendance and quality of work are positive and about 
 .40, being higher for the extreme age-groups, and less for the middle 
 group. 
 
 The r between attendance and kind of teacher is about .15, increasing 
 with each age-group, the largest jump occurring between the 5-7 and 8-11 
 groups. The r between attendance and the school plant is smaller and 
 more variable. If there is a difference between the age-groups, the r is 
 less with the oldest group than with the other two. 
 
 In three counties the r between attendance and community is appro.xi- 
 mately .30, and in the other two it is slight. So far as there is a difference, 
 it is less with the 5-7 age-group. 
 
 The r between distance and age-grade is about .20 or .25, and between 
 distance and quality of work al)out .18. The r of distance with teacher, 
 school plant, and community is very small. Age-grade relation correlated 
 w'ith quality of work gives about .25 for the first two age-groups, and 
 about .50 for the oldest children. Age-grade relation with teacher and 
 school plant is slight; with community it is a little higher, being about .15. 
 
 The r of quality of work with teacher, school plant, and community is 
 approximately 0. Quality of work is an independent variable within each 
 school and would give when correlated with these factors if the children 
 were in a single group. Two counties have correlation between teacher 
 and school plant : the other three have .25, ,32, and .38. Between teacher 
 and community the r is fairly constant and is about .35. The r between 
 school plant and conmiunity varies from in Allegany to .79 in W i- 
 comico. It is .12, .17, and .27 in the other three counties. 
 
Analysis of Significant Factors 63 
 
 Isolation of the Factors 
 
 The correlations show that children farther from school are more often 
 behind for their ages, do inferior work, and attend fewer days than chil- 
 dren living near the school ; but they do not show whether children farther 
 from school do inferior work because they live farther from school or 
 because they attend fewer days. 
 
 Again, the better teachers are more often found in the schools with 
 better buildings and equipment, which are more often found in the bet- 
 ter communities. It would be difficult to draw any conclusion in regard 
 to these three factors from Table 18, as the three things are more or less 
 tied up together. 
 
 The median r's for each age-group were taken from Table 18 as a meas- 
 ure of the central tendencies of the several relations. To get the central 
 tendency of the several relations with the factors isolated, these r's were 
 freed from their connection with each other bv the formula :^ 
 
 R, 
 
 V (l-'-^3) (i-'-^a) 
 
 By the same method all the r's in Table 18 were freed from their con- 
 nection with each other (Table 19) and the median of these partial rs was 
 taken as a second measure of the central tendency of the relations when 
 isolated. The medians of the original r's are given in Table 20. 
 
 The relatively high negative correlation between attendance and dis- 
 tance remains almost as high, indicating that the connection between 
 attendance and distance was only slightly strengthened by the joint action 
 of other factors. The same is true of the correlation of attendance with 
 age-grade and quality of work, where the connections were helped a 
 little more by operating factors. But the r's of distance with age-grade 
 and quality of work were practically all due to their connection with 
 attendance. In the original coefficients (Table 18), there is a negative 
 correlation of distance with age-grade and quality of work, because age- 
 grade and quality of work were tied up with attendance, which has a neg- 
 ative correlation with distance; but children living farther from school 
 are more often behind for their ages and more often do inferior work 
 not because they are farther away, but because they attend fewer days. 
 Being farther away has something to do with attendance, which, in turn, 
 has something to do with the age-grade and quality of work : but distance, 
 in itself, has nothing to do with progress in school or with the quality of 
 work. 
 
 ^ Yule, G. Undy, An Introduction to the Theory of Statistics, London, 1916, 
 Chap. 12, 
 
64 Factors Controlling Rural School Attendance 
 
 TABLE 19 
 
 Showing the Partial Coefficients of Correlation Obtained by Freeing 
 
 THE Original r's of Table 18 from Their Connection 
 
 with Each Other ^ 
 
 Distance Age Gr. QuaL of W. Teacher Phys. Pro. 
 
 5-7 8-11 12+5-7 8-11 12+5-7 8-11 12+5-7 8-11 12+5-7 8-11 12+ 
 
 CY 1 -.85 -.48 -.02 .34 .37 .53 .57 .29 .09 -.20 .03 .13 .39 .21 .04 
 
 CY 2 -.42 -.29 -.15 .37 .35 .39 .31 .19 .00 .04 .11 .20 -.18 -.15 -.49 
 
 CY 3 -.73 -.44 -.44 .57 .33 .24 .81 .30 .55 .41 .19 -.08 .35 -.13 -.05 
 
 CY 4 -.63 -.63 -.44 .03 .50 .27 .36 .00 .09 .51 .42 .19 .24 -.35 -.05 
 
 CY 5 -.44 -.38 -.25 .59 .75 .45 -.11 .33 .13 .20 .01 .13 -.34 -.47 -.81 
 
 CY 1 .22 .03 -.09 .45 .19 -.05 -.27 -.05 -.39 .38 .23 .07 
 
 CY 2 .04 -.16 -.03 -.09 .01 -.07 .07 .01 .22 -.03 .17 .01 
 
 CY 3 -.31 .01 -.06 .00 .08 .16 .37 .27 -.05 .26 -.22 .04 
 
 CY 4 -.05 .18 .08 .15 .01 -.09 .25 .18 .11 .18 -.28 -.09 
 
 CY 5 .18 .05 -.04 -.12 -.02 .14 .21 -.06 -.07 -.28 -.20 .00 
 
 CY 1 -.02 .10 .32-.24-.32-.12-.il .02-01 
 
 CY 2 -.12 .10 .49 -.23 -.21 -.17 .04 .21 .08 
 
 CY 3 .34 .13 .33 .13 -.19 -.01 .05-03 .17 
 
 MY 4 .10 .21 .49 .22 -.07 .00 .14 .28 -.13 
 
 CY 5 .34 -.09 .07 -.30 -.08 -.16 -.30 -.25 -.25 
 
 CY 1 .18 -.12 -.15 -.14 -.01 -.08 
 
 CY 2 -.01 .12 -.08 -.05 .01 .09 
 
 CY 3 -.39 -.01 -.01 -.38 .03 .04 
 
 CY 4 .02 -.02 -.11 -.21 -.15 .11 
 
 CT 5 .20 -.10 -.20 .15 .23 .18 
 
 CY 1 .11 .03 .01 
 
 CY 2 .26 .24 .34 
 
 CY 3 -.03 .18 .22 
 
 CY 4 .17 .18 -.05 
 
 CY 5 .18 .20 .22 
 
 The r between age-grade and quality of work is about .10 for the first 
 two age-groups and about .35 for the 12 group. Children farther along 
 for their ages more often do better work, but the difference for children 
 under 12 years of age is unimportant. 
 
 Tables 19 and 20 clearly show that children living farther from school 
 have poorer attendance, regardless of the grades they are in, the quality 
 of work they do, the kind of teachers they have, the kind of school plants, 
 or the communities in which the schools are located. 
 
 The same r's show that children behind their grades do not attend so well, 
 regardless of the distance they live from school, the quality of work they 
 do, the teachers, or the school plants they have, or their communities. 
 Likewise, those doing an inferior quality of work go fewer days regard- 
 less of anything else. The better teachers have better attendance regardless 
 
 ^The Community factor, not shown in this table, has been thrown out of the /s 
 above. 
 
Analysis of Significant Factors 65 
 
 TABLE 20 
 
 Showing the Medians of the Original Crude r's (Original M), the 
 Partials of the Same Medians (Partial M), and the Medians of 
 THE Partials of the Original r's (M Partial)* 
 
 
 Distance 
 
 
 Age Gr. 
 
 
 QuaL 
 
 of W. 
 
 Teacher 
 
 
 Phys. Pro. 
 
 Com, 
 
 
 a, 
 
 O 
 
 • 
 
 o 
 
 -.54 
 
 a 
 -.53 
 
 a 
 -.63 
 
 CO 
 
 a 
 '5i 
 
 'u 
 O 
 
 .38 
 
 .27 
 
 .37 
 
 a 
 
 o 
 
 .41 
 
 a 
 .36 
 
 OS 
 
 Hi 
 .36 
 
 .9 
 
 o 
 
 .08 
 
 P. 
 .05 
 
 CO 
 
 .20 
 
 a 
 
 O 
 .06 
 
 a 
 .10 
 
 t-i 
 
 .24 
 
 o 
 
 a 
 
 O 
 .08 
 
 .2 
 
 cS 
 
 .03 
 
 ^s-u 
 
 -.45 
 
 -.43 
 
 -.44 
 
 .50 
 
 .44 
 
 .37 
 
 .33 
 
 .25 
 
 .29 
 
 .16 
 
 .08 
 
 .11 
 
 .07 
 
 -.01 
 
 -.15 
 
 .30 
 
 .28 
 
 g 12+ 
 
 <4 
 
 -.32 
 
 -.25 
 
 -.25 
 
 .55 
 
 .35 
 
 .39 
 
 .47 
 
 .30 
 
 .09 
 
 .18 
 
 .21 
 
 .13 
 
 .00 
 
 -.09 
 
 -.05 
 
 .17 
 
 .05 
 
 g 5-7 
 
 
 
 
 -.17 
 
 .03 
 
 .04 
 
 -.17 
 
 .08 
 
 .15 
 
 .00 
 
 .05 
 
 .21 
 
 .09 
 
 .15 
 
 .18 
 
 -.02 ■ 
 
 -.01 
 
 g 8-U 
 
 
 
 
 -.20 
 
 .02 
 
 .03 
 
 -.13 
 
 .02 
 
 .01 
 
 -.10 
 
 -.03 
 
 .01 
 
 -.06 ■ 
 
 -.01 
 
 -.20 
 
 .02 
 
 .20 
 
 |12+ 
 
 
 
 
 -.20 - 
 
 -.03 
 
 -.04 
 
 -.18 
 
 -.01 - 
 
 -.05 
 
 .04 
 
 .07 
 
 -.05 
 
 .07 
 
 .03 
 
 .01 
 
 .04 
 
 .06 
 
 
 12 + 
 
 
 5-7 
 8-U 
 
 C3 
 
 12 + 
 
 t-i 
 
 .5-7 
 
 _4, 
 
 8-11 
 
 12+ 
 
 
 5-7 
 8-11 
 
 J3 
 
 12+ 
 
 .25 .12 .10 -.02 -.12 - .23 .03 .01 .04 .13 .16 
 .24 .08 .10 .01 -.13 -.19 .08 .03 .02 .08 -.04 
 .53 .37 .33 .04 -.10 -.12 -.03 -.09 -.01 .20 .20 
 
 .06 .11 .02 -.08 -.12 -.14 -.10 -.18 
 .00 -.08 -.02 .08 .06 .01 .03 -.10 
 .04 -.16 -.11 .06 .12 .09 .00 -.12 
 
 .25 .21 .18 .35 .35 
 .25 .21 .18 .35 .30 
 .25 .21 .22 .35 .29 
 
 .17 .06 
 .17 .12 
 
 .17 .12 
 
 a, 
 
 of all other factors ; but when freed from its connection with teacher 
 and community, the school plant has practically no correlation with at- 
 tendance. The r between age-grade and teacher, which was strengthened 
 originally by the relation the two factors had with community, and which 
 was practically 0, becomes about .15 when freed from its connection with 
 other factors. The fact that the better teachers classify their pupils 
 lower than other teachers, or promote them less often, suggests that the 
 
 *Children in CY 1 and CY 3 have practically no state roads or transpor- 
 tation. 
 
G6 Factors Controlling Rural School Attendance 
 
 meaning of the completion of a grade of school work varies considerably 
 in the different schools. Not only does the better teacher classify the 
 children lower according to their ages, but she marks them less accord- 
 ing to age and size, giving the younger children slightly higher marks. 
 This may mean, however, that the better teachers do a little more for 
 younger children. 
 
 The greatest factor in the school attendance of the country child in 
 these five counties is the distance he lives from school. Next in order 
 is the relation of his grade to his age ; and about equal to the age-grade 
 relation is the quality of his school work. Next in order come the teacher 
 and the community, which are similar in importance, differences between 
 teachers being a little more significant than differences between the ccmi- 
 munities. 
 
 Causal Relations 
 
 A coefficient of correlation shows the extent to which one fact varies 
 with another. It does not show which one is the cause of the other, or 
 whether the variation of the two is not controlled by a third factor. In the 
 case of poor attendance and poor quality of work, both factors are per- 
 haps both cause and effect. Poor attendance causes a poor quality of work, 
 because attendance is the only fact with which quality of work has an im- 
 portant independent connection. But since a child behind the proper grade 
 for his age attends fewer days than one up to grade, it is very probable 
 that a child, unsuccessful in doing his school tasks and at the foot of the 
 class with which he recites, would attend fewer days because of the un- 
 satisfactory character of his school work. 
 
 The grade in which a child is enrolled, having been determined before 
 the opening of school in the fall, could not be the result of attendance dur- 
 ing a year following, after the age-grade relation had been determined. 
 The age-grade relation must be the cause of the attendance so far as a 
 causal relation exists between these two facts. 
 
 Likewise, the attendance of the children during any year could not be 
 the cause of the kind of teachers or community they have or the dis- 
 tances they live from school. These were determined before the school 
 year opened. The kind of attendance is the result and not the cause of 
 these facts. 
 

 The Analysis of Significant Faclors 
 ^ ? 9 JC 
 
 + fi, 00 »o 
 
 fo 1—1 
 
 iDis^o fMrs.po i^oo'* t^r^io 
 
 67 
 
 E E = 
 
 -' "-' rt 
 •5»-0"^'-^ po^s.\o vO'<t'^ oor^r~^ (NifviOs on^ii 
 
 <<rg<~Of*5 ,-<<M,-i -H-^i-co f^r^-^ (Nl'<4-CM "-l-Ug 
 
 < p 
 
 ^55S 
 
 ^fT 00r^<^ OOCg^O rovO<^ fMTj-T}- (MVOfVl 
 l*^ Ogro OvirO ,— I,— (,_(— Hr— .— ifM 
 
 
 OvOOfO fOO"^ \OOs--i 
 
 ^ 5 
 
 R Q 
 
 ^o rs. CO 00 
 
 S^ CM ro TJ- 
 
 .-1 , CO 00 Tj- <-^ ro 
 
 O 0\ O ■* t>. 00 
 
 3 lu -a 
 
 ^ -o 
 
 It: 
 
 ►J 
 
 _, r^ •-^VOOO rOt^vCNJ IDIOOO VDOrO 
 
 J?; •«*■ 00 vo 00 -^ (M 
 
 
 o a 
 g E 
 
 — OC5 •^(^■^ O'OOV O^-"^ ^^0 0^ 
 
 W J ^o 0\ O \0 
 
 ^ ^ ^ '^ CM CO 
 
 Q^ r-iOTj-CN OfM-* ^hO\00 .—100"^ 0x0"-- 
 
 S ^ ^ 
 
 c ii. i^_^^o— ' ^vO"-^ ^Cvi'^ ^^o-^ '0'^ir~^ 
 
 sL ij Zoor^oo >o<~oro r^ivO'T c^vo^o •<^o"^ 
 
 y. 
 
 be 
 < 
 
 a^ rs <-• 
 
 t:; « •/ 
 
 
 
 
 
 
 
 "t; 
 
 
 
 
 
 
 5 
 
 
 
 
 
 
 C (/) 
 
 > 
 
 
 
 
 
 
 u *- 
 
 
 
 
 
 
 
 a, u 
 
 u- 
 
 tn 
 
 to 
 
 
 U3 
 
 U3 
 
 u o 
 
 O 
 
 3 
 
 3 
 
 
 3 
 
 3 
 
 I- Q. 
 
 
 
 
 
 
 QJ 
 
 
 r^::: ^ 
 
 r^:^ «■ 
 
 r^ ^ 
 
 "H, 
 
 r^::: «« 
 
 o a 
 
 i^' 
 
 u^oi::^ 
 
 U^oi,^ 
 
 J,oi> 
 
 CM 
 
 u!,oi^ 
 
 c 
 
 "« 
 
 if} ■— " 
 
 <. 
 
 A 
 
 I_J 
 
 
 :^. 
 
 J 
 
 
 ,< 
 
 ■A 
 
 Ui 
 
 3 
 
 O >, 
 
 u 
 
 u 
 
 u 
 
 G «> rt o 
 
 >< "E bf, S 
 8.? := 
 
(58 factors Controlling Rural School Attendance 
 
 TABLE 22 
 Showing Median Days Attended by Children Living at Different 
 
 Distances from School, by Age-Groups and Counties, with Dis- 
 tance Uncorrected for Kind of Roads and Days Transported 
 
 Age- Distance in Miles 
 
 County Groi'i- -% 
 
 CY 1 5-7 145 
 
 8-11 151 
 
 12+ 113 
 
 CY 2 5-7 149 
 
 8-11 148 
 
 124- 125 
 
 CY 3 . 5-7 137 
 
 " 8-11 144 
 
 12+ 123 
 
 CY 4 5-7 170 
 
 8-11 172 
 
 12+ 150 
 
 CY 5 5-7 139 
 
 8-11 148 
 
 12+ 112 
 
 TABLE 23 
 Showing the Days Attended by Mediant Pupil Living at Different Dis- 
 tances, after Distance was Corrected for Kind of Roads and Days 
 Transported 
 
 County Group -Yz -1 Over 1 
 
 CY 1 
 
 CY 2 
 
 -/2 
 
 -1 
 
 -IK2 
 
 _2 
 
 Over 
 
 132 
 
 99 
 
 95 
 
 87 
 
 70 
 
 130 
 
 119 
 
 126 
 
 79 
 
 82 
 
 75 
 
 75 
 
 69 
 
 75 
 
 44 
 
 125 
 
 128 
 
 115 
 
 110 
 
 90 
 
 135 
 
 129 
 
 134 
 
 114 
 
 99 
 
 99 
 
 87 
 
 81 
 
 78 
 
 58 
 
 89 
 
 119 
 
 79 
 
 70 
 
 85 
 
 135 
 
 133 
 
 105 
 
 113 
 
 112 
 
 125 
 
 118 
 
 77 
 
 79 
 
 75 
 
 163 
 
 161 
 
 157 
 
 140 
 
 143 
 
 174 
 
 152 
 
 142 
 
 150 
 
 149 
 
 140 
 
 132 
 
 116 
 
 115 
 
 100 
 
 139 
 
 127 
 
 124 
 
 103 
 
 119 
 
 136 
 
 128 
 
 115 
 
 116 
 
 115 
 
 90 
 
 94 
 
 91 
 
 87 
 
 79 
 
 CY 3 
 
 CY 4 
 
 CY 5 
 
 Age- 
 
 
 Distance in Miles 
 
 
 Group 
 
 -Y 
 
 -1 
 
 Over 
 
 5-7 
 
 139 
 
 104 
 
 72 
 
 8-11 
 
 146 
 
 125 
 
 94 
 
 12 plus 
 
 91 
 
 71 
 
 68 
 
 5-7 
 
 146 
 
 124 
 
 108 
 
 8-11 
 
 146 
 
 134 
 
 115 
 
 12 plus 
 
 118 
 
 87 
 
 7Z 
 
 5-7 
 
 131 
 
 120 
 
 85 
 
 5-11 
 
 142 
 
 133 
 
 113 
 
 12 plus 
 
 124 
 
 118 
 
 76 
 
 5-7 
 
 168 
 
 162 
 
 130 
 
 8-11 
 
 169 
 
 156 
 
 139 
 
 12 
 
 147 
 
 145 
 
 96 
 
 5-7 
 
 139 
 
 127 
 
 111 
 
 8-11 
 
 142 
 
 129 
 
 106 
 
 12 plus 
 
 108 
 
 95 
 
 85 
 
Analysis of Significant Factors G9 
 
 The zero correlation of distance with age-grade and quality of work 
 shows that distance is not a selective factor. Children do not vary in 
 ability according to the distances they live from school. They vary only 
 in achievement, which is the result of the variation of attendance with 
 distance. 
 
 Summary 
 
 1. The age-grade relation of each child was scored by assigning values 
 according to the per cent of the children of his age that came between 
 his age-grade group and the median of all the children of that age. 
 
 2. The quality of work of each child was scored by assigning values ac- 
 cording to the per cent of the children in the school that was between 
 the quality-of-work group with which the child was marked and the 
 median of all the children in his school. 
 
 3. The teachers were scored by combining marks for their training, salary, 
 and rating; the school plants were scored by combining marks for the 
 rating of the building, grounds, and equipment. 
 
 4. The distance of each child was corrected for kind of roads and days 
 transported. 
 
 5. The correlation between the factors was computed by the method of 
 unlike-signed pairs and the coefficients isolated by partial correlation. 
 
 6. The ;- between distance from school and attendance is negative, and 
 higher than the other coefficients ; the r's of attendance with the other 
 factors are positive, and come in the following order : age-grade rela- 
 tion, quality of work, teacher, and community. 
 
■xm 
 
 021 731 563 3 
 
 
 ' i' "I ' 
 
 a 
 
 ^^^^.;'' 
 
 'W';<