151t.yv T~ JJIAYVERITAS O h F TI 72 ITYJFjlco, ILLITERACY IN THE UNITED STATES SEE LAST PAGES IN THIS BOOK FOR A COMPLETE LIST OF THE UNIVERSITY OF NORTH CAROLINA SOCIAL STUDY SERIES The University of North Carolina Press, Chapel Hill, N. C.; The Baker and Taylor Company, New York; Oxford University Press, London; Maruzen-Kabushiki-Kaisha, Tokyo; Edward Evans s Sons, Ltd., Shanghai. ILLITERACY IN THE UNITED STATES BY SANFORD WINSTON Associate Professor of Sociology North Carolina State College;, LIBI4YrAS. P- * * * *: * * CHAPEL HILL THE UNIVERSITY OF NORTH CAROLINA PRESS 1930 COPYRIGHT, 1930, BY THE UNIVERSITY OF NORTH CAROLINA PRESS * * PRINTED IN THE UNITED STATES OF AMERICA BY EDWARDS & BROUGHTON COMPANY, RALEIGH N. C., TO ELLEN HAL -,,.2, PREFACE The role of education in equipping the individual for adjustment to the complexities of modern life grows increasingly significant. The recent efforts in the United States to eliminate illiteracy are largely in recognition of this fact. Because of these factors, it is believed that the timeliness of the subject warrants the publication of this book instead of holding up publication until additional data are forthcoming. The trends discussed, as well as the relationships analyzed, are believed to be true in so far as results based upon present data may be considered true. The first objective, therefore, was to analyze the trend of illiteracy in the United States and its present relation to sex, age, urban and rural environment, race and nationality, and school systems. The second purpose was to emphasize the fact that illiteracy, as a measure of educational status, achieves importance as it affects other social phenomena. Its relationship with the selected factors of birthrate, infant mortality, early age of marriage, size of family, mobility, suicide, and urbanization has been quantitatively determined. All the results of the statistical techniques utilized have been presented with a definite knowledge of their limitations. Social data are exceedingly complex and the most careful methods of refining them still leave many factors uncontrolled. Each factor studied was selected because its relationship to illiteracy could be definitely posited on an empirical basis. In no case are the results to be regarded as exact measurements but rather as probabilities on the basis of which concrete inferences are possible. X PREFACE All data and the computations based upon them have been computed with the greatest care and thoroughly checked. In a study involving such a number of complicated calculations, however, errors can scarcely be excluded. If any such are found, the author is responsible, although every effort has been made to reduce them to a minimum. The methodology followed is included in Appendix A. The author wishes to acknowledge his indebtedness for unvarying stimulation and keen criticism to Professor Pitirim Sorokin; for various suggestions and criticisms to Professors R. W. Murchie, E. H. Sutherland, W. F. Ogburn, and Dean Carl C. Taylor; for advice on certain technical questions to Professors Henry Schultz, F. C. Mills, G. W. Foerster, and Bruce D. Mudgett, Dr. Mordecai Ezekiel, and especially to Professor Marc Leager; for constant stimulation and assistance in many details to Ellen Winston. Finally a word of appreciation is due Professor Howard W. Odum for his cooperation in the furtherance of the study's publication. S.W. CONTENTS PART I THE GENERAL PROBLEM OF ILLITERACY IN THE UNITED STATES Chapter Page I. SCOPE AND METHOD OF THE STUDY 3 II. ILLITERACY IN THE UNITED STATES, I870 TO I920o -......-. —. ---.-. — --- 7 III. ILLITERACY AND SEX.......... --- —----. 28 IV. ILLITERACY AND AGE........ --- —----- 41 V. ILLITERACY AND URBAN AND RURAL ENVIRONMENT T..... —. ---.-. 48 VI. ILLITERACY AND RACE AND NATIONALITY ------.. --- —---—.. 57 VII. ILLITERACY AND SCHOOL SYSTEMS 68 VIII. ADJUSTED ILLITERACY RATES ---... 73 IX. CONCLUSIONS TO PART I..-........ --- 79 PART II ILLITERACY IN RELATION TO CERTAIN SOCIAL PHENOMENA X. THE RELATION OF ILLITERACY TO BIRTH-RATE.-........... —.. ------ 85 XI. THE RELATION OF ILLITERACY TO INFANT MORTALITY..-............ ---- 94 XII. THE RELATION OF ILLITERACY TO EARLY AGE OF MARRIAGE --- —-- IOI xii CONTENTS Chapter Page XIII. THE RELATION OF ILLITERACY TO SIZE OF FAMILY ---...... —........-...-..-. o19 XIV. THE RELATION OF ILLITERACY TO MOBILITY TO OTHER STATES —. I I7 XV. THE RELATION OF ILLITERACY TO SUICIDE___.......... --- —--—.-. ---.- 125 XVI. THE RELATION OF ILLITERACY TO URBANIZATION AND TO SCHOOL SYSTEMS. --- —--------------—. --- —- 132 XVII. CONCLUSIONS TO PART II ----------- 140 APPENDIXES APPENDIX A, METHODOLOGY.. -------------- I49 APPENDIX B, SUPPLEMENTARY TABLES —. I57 BIBLIOGRAPHY. --- —-........ --- —------- 163 PART I THE GENERAL PROBLEM OF ILLITERACY IN THE UNITED STATES I CHAPTER I SCOPE AND METHOD OF THE STUDY In presenting this investigation of illiteracy in the United States and its relation to other social factors the study has been divided into two parts. Part I deals with the general problem of illiteracy as it exists in the United States at the present time, together with the trend of illiteracy for the past fifty years, and a brief summary of the data on illiteracy prior to I870. The connotations in regard to sex, age, urban and rural environment, race and nationality, and school systems are also investigated in some detail. Finally, the method of standard population is utilized to present a corrected alignment of the forty-eight states when the native whites of native parentage alone are considered. Part II studies illiteracy in its relation to certain selected social phenomena. These latter include birth-rate, early marriage, infant mortality, size of family, urbanization, mobility to other states, and suicide. Simple, partial, and multiple correlations have been utilized to study the various relations while the regression equations offer a basis for prediction. There are five earlier studies on illiteracy in the United States of significance. The first of these, by Edwin Leigh, was published in the Report of the Commissioner of Education for I8701 and consisted of a thorough revision of the data of the 1840, I850, and I860 censuses. This was followed by Charles Warren's study, "Illiteracy in the United States in 1870 and I880."2 In this he has made a distinct 1Annual Report of the Commissioner of Education, 1870, pp. 467-502. 2U. S. Bureau of Education, Circular of Information No. 3, 1884. 4 ILLITERACY IN THE UNITED STATES contribution in adequately presenting in greater detail and depth of analysis the results of the I870 and I880 census reports. In I905, a bulletin of the Bureau of the Census, "Illiteracy in the United States, I905,"3 was prepared under the supervision of W. F. Willcox. In it comparisons of the results of the 900o Census with those of the two preceding Censuses were made. For the first time, the data were sufficiently accurate to permit reliable detailed comparison with recent data. The fourth study is a bulletin published by the United States Bureau of Education in 1913, "Illiteracy in the United States and an Experiment for its Elimination."4 It is chiefly a summary of the I9IO Census data on illiteracy, the second part containing a study of the Rowan County, Kentucky, experiment for the elimination of illiteracy. The fifth study, "Adult Illiteracy," by W. Talbot, was published in I9I6.5 It deals chiefly with the data on the illiteracy of immigrants from I900 to I914, showing that the most illiterate immigrants came from Mexico and from southern and eastern Europe, and the least illiterate immigrants from northwestern Europe. There are various reasons for studying illiteracy. First, illiteracy is important because it is a form of isolation. The illiterate is unable to communicate with his fellows in written symbols and hence is largely restricted to his immediate social groups for many forms of social stimuli. Illiteracy is a definite form of individual isolation. Second, modern culture is dependent upon written symbols to an important extent. Third, the illiterate person is handicapped in his reactions to stimuli and situations, to the extent that these phenomena utilize the written word. More fundamental than this, however, is the fact that 3Bureau of the Census, Bulletin No. 26, 1905. 4Bulletin No. 20. 6U. S. Bureau of Education, Bulletin No. 35, 1916. SCOPE AND METHOD OF THE STUDY 5 his range of stimuli is limited to those stimuli which are not conveyed by means of written symbols. Fourth, illiteracy is one of the factors in societal phenomena. Up to the present time the importance of its relations to other phenomena has not been adequately studied. The definition of illiteracy which is basic to this investigation is that adopted by the Department of the Census of the United States and by most foreign countries. "Illiteracy, as defined by the Census Bureau, signifies inability to write in any language, not necessarily English, regard; less of ability to read.... In general, the illiterate population as shown by the census figures should be understood as comprising only those persons who have had no education whatever. Thus the statistics do not show directly or definitely the proportion of the population which may be termed illiterate when the word is used to imply lack of ability to read and write with a reasonable degree of facility; but they do afford a fairly reliable measure of the effect of the improvement in educational opportunities from decade to decade."6 This definition introduces the important point that the illiterates are at one end of a frequency distribution set up on a basis of education. There is a high percentage of individuals in the population who, while not classified as illiterates, lack facility in reading and writing. Beyond these are the various gradations in education. The illiteracy data are presented in such form that they can be quantitatively studied. The reliability of the data utilized is adequately discussed in the census report. There is undoubtedly a margin of error in the statistics of illiteracy, resulting from a variety of causes. In some cases there may be unwillingness to admit illiteracy on the 6Fourteenth Census of the United States, vol. II, p. 1145. 2 6 ILLITERACY IN THE UNITED STATES part of persons enumerated. Furthermore, in parts of the country where practically all native white persons are literate the enumerators are likely to acquire the habit of returning them as such without the formality of an inquiry, and in this way a few isolated cases of illiteracy may be overlooked. On the other hand, in the case of Negroes the opposite assumption may sometimes be made by white enumerators, while, in the case of the foreign born, inability to write in English may sometimes be taken as constituting illiteracy, although the instructions make it clear that a person able to write in any language is to be returned as literate. For the United States as a whole and for the states and large cities the figures are probably nearly enough accurate to supply a sound basis for judgment as to the relative illiteracy of different classes of the population, of persons in different age groups, and of males as compared with females. Beyond question comparisons between different censuses show the general tendencies with substantial accuracy. The returns for small areas, however, may be open to question in some cases.7 The adequacy of the supplementary data will be discussed as the data are presented. For the purposes of this study, the census data have been taken only for the past fifty years. Before that time the validity of the returns is questionable. A brief summary of the earlier years is included for historical purposes. Data from 1870 to I920 are used to present general tendencies while detailed analysis is largely limited to the census of 1920 in which the illiteracy data are more adequately presented than in any preceding census. Hence the limitations of the analysis of the data are largely dependent upon the census organization of the material, although reorganizations of the data have been made in many instances.8 7Loc. cit. 8The data are presented for "all classes," ten years of age and over, unless otherwise specified. CHAPTER II ILLITERACY IN THE UNITED STATES, I870 TO I920 Before beginning an analysis of the data on illiteracy in the United States from I870 to I920, a brief summary of the earlier census reports on illiteracy may well be included for purposes of completeness. The first census data with reference to illiteracy in the United States were gathered in the I840 census. At that time, an enumeration was made of the number of white persons over twenty years of age who could neither read nor write. The data were published by states and territories, and by counties, and subdivisions of counties.1 In I850 classification was first made. The census for that year gave the ratio of white illiterate to total white, the ratio of illiterate native white over ten years of age to total native white, the ratio of free colored (Negro?) illiterate to total free colored, and the ratio of foreign illiterate to total foreign over twenty years of age, "supposing the (foreign) illiterate to be all white." This information was given for the geographical divisions, the thirty-one states, four territories, and the District of Columbia, and for the slave states and free states. The ratio of white illiterate to white population for I850 was compared with the same data for I840 for the six geographical divisions, for the slaveholding states as compared with the non-slaveholding states as well as for the country as a whole.2 Illiteracy was 'Sixth Census of the United States. 'Compendium of the Seventh Census, Part III, pp. 152, 153, Tables CLV, CLVI, CLVII. 8 ILLITERACY IN THE UNITED STATES here considered, as in i840, as the inability to read and write. In addition, data as to the number of whites and free colored, by totals and by sex, as well as the number of native and foreign-born, over twenty years of age who could not read or write, were given.- The opinion of Superintendent J. D. B. DeBow was that the statistics were "generally accurate."' The i86o Census tabulated the number of whites and free colored who could not read and write, by totals and by sex. A division into natives and foreign (no sex division being made) was also given. The data were tabulated for the thirty-four states, six territories, and the District of Columbia. No returns were received from the territory of Colorado.5 In i870, the census enumerators for the first time classified as illiterate all persons ten years of age and over who could not write in any language. Hence the data became for the first time comparable with succeeding census reports. Futhermore, the census stated that, great numbers of persons rather than admit their ignorance, will claim to read, who will not pretend that they can write.... If a man cannot write, it is fair to assume that he cannot read well; that is, that he really comes within the illiterate class.... Taking the whole country together, hundreds of thousands of persons appear in the class "cannot write" over and above those who confess that they cannot read. This is the true number of the illiterate of the country.6 Thus the data became not only comparable but also sufficiently reliable, in those categories utilized in this study. 'Seventh Census of the United States, p. LXI, Table XLIII. 4See p. 10, Introductory Remarks, Ibid., 5Eighth Census of the United States, vol. IV, p. 508. 6Ninth Census, Population and Social Statistics, XXX. FROM 1870 To 1920 9 The trend of illiteracy in the United States has been steadily downward.7 The relative decrease is from 20.0 per v cent illiteracy in 1870 to 6.0 per cent in I920. Expressed in another way, one out of every five persons ten years of age and over was classed as illiterate in 1870, while in 1920, approximately one person in every seventeen was classified as illiterate. The number of illiterates in I920 (4,931,905) was approximately three-quarters of a million less than the number in 1870 (5, 658,144), the high-water mark being reached in I890, when the number rose to 6,324,702. Figure I shows the illiteracy rates for the years 1870 to 1920, which the trend line follows closely. The actual data TABLE I ILLITERACY IN THE UNITED STATES, 1870 TO 1920 Yr Number of Persons 10 Years Number of Per Cent of Age and Over Illiteratess Illiterate 1870 28,228,945 5,658,144 20.0 1880 36,761,607 6,239,958 17.0 1890 47,413,559* 6,324,702* 13.3 1900 57,949,824 6,180,069 10.7 1910 71,580,270 5,516,163 7.7 1920 82,739,315 4,931,905 6.0 *Exclusive of persons in Indian Territory and on Indian Reservations. Such areas were especially enumerated in 1890, but illiteracy statistics are not available. indicate a slower rate of decrease in the future, however, than is presented by the trend. The United States Census Bureau classifies the various states into nine geographical divisions.9 Illiteracy in the 7See Table I. 8Fourteenth Census of the United States, vol. II, chap. XII, Table I. 9New England, Middle Atlantic, East North Central, West North Central, South Atlantic, East South Central, West South Central, Mountain, an dPacific. & ~ ~ 7 I9f st TO-19t FROM 1870 TO 1920 11 first of these, the New England Division, is increasing in numbers, but decreasing in percentage (Table II). Trend lines fitted to the data for the six New England states show a steady decrease in Rhode Island with slight decreases in Vermont and Massachusetts. Maine and Connecticut reveal but little decrease in the trend, while New Hampshire shows an increase for the total period. By observation, it is evident that the trend since i890 is downward. The slowness of the decrease in the illiteracy percentage is partly a function of a relatively low illiteracy rate in i870, and partly a result of the influx of illiterate immigrants, particularly since i88o. This latter factor will be discussed in a later chapter. The trend for the Middle Atlantic Division as a whole, from i870 to I920, is clearly defined."0 The trend lines for New Jersey and Pennsylvania show a slight though steady decrease. The illiteracy rate for New York, -which was 7.1 per cent in i870, dropped to 5.5 per cent in i88o, and remained at this level until the last census, when the downward trend was resumed. Only in i88o however, did the actual number of illiterates decrease. The East North Central Division, comprising the states of Ohio, Indiana, Illinois, Michigan, and Wisconsin, shows a gradual decrease, from an illiteracy rate of 8.2 per cent, in I870, to a rate of 2.9 per cent in 1920, with the exception of i890 when the illiteracy decrease slowed up, and the rate actually increased in Michigan and Wisconsin. Of the three divisions so far discussed, the East North Central Division is the only one to show a decrease in the actual-number of illiterates in the half-century span. From a I2.0 per cent illiteracy rate in i870, the West North Central Division, comprising the states of Minne "0See Table II and Figure 3. 12 ILLITERACY IN THE UNITED STATES TABLE II ILLITERACY IN THE UNITED STATES, BY DIVISIONS, 1870 TO 192011 Number of Section Year Persons 10 Number of Per Cent Years of Age Illiterates Illiterate and over 1870 New England............. Middle Atlantic........... East North Central........ West North Central....... South Atlantic............ 1870 1880 1890 1900 1910 1920 1870 1880 1890 1900 1910 1920 1870 1880 1890 1900 1910 1920 1870 1880 1890 1900 1910 1920 1870 1880 1890 1900 1910 1920 1870 1880 1890 1900 1910 1920 2,773,337 3,219,856 3,859,728 4,524,602 5,330,914 5,945,989 6,657,455 8,050,234 10,028,649 12,167,559 15,446,515 17,666,354 6,586,383 8,339,175 10,317,783 12,443,302 14,568,949 17,130,786 2,706,051 4,421,666 6,591,830 7-,838,564 9,097,311 9,889,740 4,207,398 5,286,645 6,415,921 7,616,159 9,012,826 10,513,447 3,109,016 3,831,101 4,608,235 5,474,227 6,178,578 6,677,229 195,963 198,506 243,404 272,402 280,806 289,700 516,314 500,863 616,585 704,134 873,812 865,832 542,448 507,286 588,965 534,299 491,850 495,470 323,469 345,734 373,303 324,023 263,138 193,221 1,943,166 2,129,830 1,981,888 1,821,346 1,444,294 1,212,942 1,393,195 1,565,762 1,433,669 1,364,935 1,072,100 845,459 7.1 6.2 6.3 6.0 5.3 4.9 7.8 6.2 6.2 5.8 5.7 4.9 8.2 6.1 5.7 4.3 3.4 2.9 12.0 7.8 5.7 4.1 2.9 2.0 46.2 40.3 30.9 23.9 16.0 11.5 44.8 40.9 31.1 24.9 17.4 12.7 East South Central........ FROM 1870 TO 1920 13 TABLE II-Continued Number of Section Year Persons 10 Number of Per Cent Years of Age Illiterates Illiterate and over West South Central....... 1870 1,439,204 631,200 43.9 1880 2,245,142 836,827 37.3 1890 3,191,252 885,202 27. 7 1900 4,649,988 953,644 20.5 1910 6,394,043 845,604 13.2 1920 7,739,536 773,637 10.0 Mountain................ 1870 237,638 74,939 31.5 1880 500,441 90,408 18.1 1890 890,252 101,903 11.5 1900 1,276,076 122,901 9.6 1910 2,054,249 140,737 6.9 1920 2,564,463 132,659 5.2 Pacific................... 1870 512,463 36,450 7.1 1880 867,347 64,742 7.5 1890 1,509,909 97,783 6.5 1900 1,959,347 82,385 4.2 1910 3,496,885 103,822 3.0 1920 4,611,771 123,435 2.7 sota, Iowa, Missouri, North Dakota, South Dakota, Nebraska, and Kansas, achieved an illiteracy rate of 2.0 per cent in I920, the lowest attained by any of the nine divisions in the United States. The South Atlantic Division comprises the states of Delaware, Maryland, Virginia, West Virginia, North Carolina, South Carolina, Georgia, and Florida, as well as the District of Columbia.12 The high percentage of "Fourteenth Census, vol. II, chap. XII, Table V; Thirteenth Census, vol. I, chap. XIII, Table XXVI; Thirteenth Census, vol. I, chap. IV, Table XLIII (data computed); Compendium of the Eleventh Census, Part III, pp. 301-2 (data computed); Ninth Census, vol. I, Tables IX and XXVI (data computed). 12In general, data for the District of Columbia are omitted as it is not a comparable area. 14 ILLITERACY IN THE UNITED STATES illiteracy in I870, 46.2 per cent, is partly due to the presence of a large Negro population, partly to the economic and social disorganization in most of this section in the preceding decade, partly to the rural conditions, and in large part to the less strongly entrenched cultural traitcomplex of institutional education. Virginia, North Carolina, South Carolina, Georgia, and Florida had an illiteracy rate of over 50.o per cent in I870. The marginal states of Delaware, Maryland, and West Virginia had an illiteracy rate of approximately half that of the other states. The tremendous sweep of the downward trend lines shows the progress being achieved by all the South Atlantic States. The states of Kentucky, Tennessee, Alabama, and Mississippi compose the East South Central Division. The illiteracy trend is fundamentally the same as that of the South Atlantic Division. The plotted data (Figure 7) reveal that the illiteracy trends dropped more rapidly after I880 than before in each of the four states. This, as in the preceding division, is in large part a function of the greater disorganization and correspondingly greater recovery in the second as compared with the first decade after the outbreak of the Civil War. The States of Arkansas, Louisiana,. Oklahoma, and Texas comprise the West South Central Division. These states, the western marginal states of the South Eastern area of the United States, reveal high rates of illiteracy with the exception of Oklahoma. The comparatively low illiteracy of this state is due in part to the lower percentage of Negroes in the population.13 The rise in 900o for Okla"Negroes formed the following percentages of the population of the states of this division in 1920: Arkansas 27.0 Louisiana 38.9 Oklahoma 7.4 Texas 15.9 FROM 1870 TO 1920 15 homa, is due to the inclusion of Indians who were not included in i89o.14 The rapid progress of the Mountain Division is reflected in the drop from 3I.5 per cent illiterate in i870 to 5.2 per cent in i920. The early settlement conditions and the large percentage of illiterate Mexicans and pre-literate American Indians probably account for much of the early high percentage of illiteracy.15 The rate of illiteracy on the Pacific Coast has never been high, comparatively speaking, while since i88o in the case of Oregon and Washington, and since i890 in California, the trend has been slowly downward. The steady increase in the number of illiterates in the Pacific Division is due in part to the high rate of illiteracy of the Japanese and Chinese immigrants, but chiefly to the Mexican and Mexican-descendant immigration.1 Much of this immigration has used California as the state of entrance, hence its relatively larger proportion of illiteracy. In conclusion, it may be stated that the actual rise in rate of illiteracy for many of the states in i890 or a slowing up of the downward trend in a still greater number in that year may be chiefly ascribed to the high percentage of illiterates among the "new " immigration from southern and eastern Europe. This phenomenon, however, does not negate the general statement that the trend of illiteracy for the United States and its divisions, as well as for the forty-eight states, has been steadily downward. "4Persons in Indian Territory and on Indian Reservations were especially enumerated in 1890, but illiteracy statistics are not available. '5The data apparently present a number of inconsistencies which are symptoms of inaccurate census returns for some states prior to 1900. "8For proof of this, see chap. VI. 16 ILLITERACY IN THE UNITED STATES TABLE III PERCENTAGES OF ILLITERACY IN THE UNITED STATES, BY STATES 1870 TO 192017 States 1870 1880 1890 1900 1910 1920 Maine........................ 3. 8 4.3 5.5 5.1 4. 1 3.3 New Hampshire................ 3.8 5.0 6.8 6.2 4.6 4.4 Vermont...................... 6.8 6.0 6. 7 5. 8 3.7 3.0 Massachusetts................. 8.4 6.5 6.2 5.9 5.2 4. 7 Rhode Island................ 12.6 11.2 9.8 8.4 7.7 6. 5 Connecticut................... 7.0 5. 7 5.3 5.9 6.0 6. 2 New York.................... 7. 1 5.5 5.5 5.5 5.5 5. 1 New Jersey.................... 8.0 6.2 6.5 5.9 5.6 5. 1 Pennsylvania................. 8. 6 7. 1 6. 8 6. 1 5.9 4. 6 Ohio.......................... 8.9 5.5 5.2 4. 0 3.2 2. 8 Indiana.................... 10.6 7.5 6.3 4.6 3.1 2.2 Illinois........................ 7.4 6.4 5. 2 4. 2 3. 7 3.4 Michigan...................... 6.1 5.2 5.9 4.2 3.3 3.0 Wisconsin..................... 7.4 5.8 6. 7 4.7 3.2 2.4 Minnesota..................... 8.0 6.2 6.0 4.1 3.0 1. 8 Iowa...................... 5.5 3.9 3.6 2.3 1.7 1.1 Missouri.................... 18.4 13.4 9. 1 6.4 4.3 3.0 North Dakota............... 14. 4. 6.0 5.6 3.1 2.1 South Dakota............... 4.2 5.0 2.9 1.7 Nebraska.................... 5. 5 3.6 3. 1 2.3 1.9 1.4 Kansas.................... 9.5 5.6 4.0 2.9 2.2 1.6 Delaware................... 24.9 17.5 14.3 12.0 8.1 5.9 Maryland.................... 23.5 19,3 15.7 11. 1 7.2 5.6 Virginia...................... 50.1 40.6 30.2 22.9 15.2 11.2 West Virginia................. 24 19.9 14,4 11.4 8.3 6.4 North Carolina................. 51.7 48.3 35.7 28.7 18.5 13.1 South Carolina................ 57. 6 55.4 45.0 35.9 25. 7 18.1 Georgia...................... 56.6 49.9 39.8 30.5 20.7 15.3 Florida........................ 54. 8 43.4 27. 8 21.9 13. 8 9.6 Kentucky.................... 35.7 29.9 21-.6 16.5 12.1 8.4 Tennessee.................... 40.9 38.7 26.6 20.7 13.6 10.3 Alabama...................... 54.2 50.9 41.0 34.0 22.9 16.1 Mississippi.................... 53.9 49,5 40.0 32.0 22.4 17.2 FROM 1870 TO 1920 17 TABLE Ill-Continued States 1870 1780 1890 1900 1910 1920 Arkansas...................... 30.2 38.0 26.6 20.4 12.6 9.4 Louisiana...................... 52.5 49.1 45.8 38.5 29.0 21.9 Oklahoma............................ 5.4 12.1 5.6 3.8 Texas..................... 37.1 29.7 19.7 14.5 9.9 8.3 Montana..................... 5. 1 5.3 5.5 6. 1 4.8 2.3 Idaho..................... 25.7 7. 1 5.1 4.6 2.2 1.5 Wyoming..................... 7. 5 3.4 3.4 4.0 3.3 2.1 Colorado...................... 22.5 6.6 5.2 4.2 3.7 3.2 New Mexico.................. 78.6 65.0 44.5 33.2 20.2 15.6 Arizona....................... 33.4 17.7 23.4 29.0 20.9 15.3 Utah....................... 13.0 9.1 5.6 3. 1 2.5 1.9 Nevada.................... 2.4 8.0 12.8 13.3 6. 7 5.9 Washington.................... 7.5 7.0 4.3 3. 1 2.0 1.7 Oregon....................... 6.8 5.7 4.1 3.3 1.9 1.5 California...................... 7.3 7.8 7.7 4.8 3.7 3.3 ' - 17Fourteenth Census, vol. II, chap. XII, Table VII; Compendium of the Eleventh Census, Part III, pp. 301, 302; Ninth Census, vol. I, Tables IX, XXVI. -01 I IT 5- K2T90 V off - ILI OZ'61 OL 8/ _-so Z4 /7 7771d4(11 lv7:~ 0 - - - -4-.111 -- _il / -I / -V N N 0) I N N o 1 — 1 -/ - I _ T - I / - / - {e L~t ac n ~Z l A&N it-Z? 2,2172al -5tr - -17 - 4~. 8 -0 - -92 0- 0 - - - -- //I I i; i I I// O / I I I// r 0^ I I I I /[t / I I I I I Y/ PI /I I I I I c 5 X/ / pk o --— 9 7 Y h i-e 7e 7, /e y/? e JOUt1~, qAA'.77t'CS e /a70 /9Z - - - - -40- -A.' -3,0 — -- --- -- -a. /5*,~, a N 0 ~ / 0 0 /ir I T T [ - 4 --- -7 - - 0 N 4 --- -I --- 4K4 N / F,9,re 3.-.////7'-a/ 7 h e Mouv","a i4 Stl~e / 70,92 0 r40' - 3 - A-Is -'4 — -8r./o 80/0 901A )r r go/9 W a0AA - -- - t I 13 / C -4 - -4 --- — - I N N 4L C 0 N. -1~ C -II, / — 4 1.il___ __I- C C 4. CHAPTER III ILLITERACY AND SEX The objective of this chapter is the analysis of the data on illiteracy and sex. The sources of error are, by the nature of the data, almost negligible when sex alone is considered. Table IV shows the number and percentage of both male and female illiterates for the United States as a whole. The folkways in reference to female education in the middle of the nineteenth century are reflected in the larger percentage of female illiteracy from 1870 to 1910. The trend is downward at a more rapid rate for females than for males, however, resulting in a slightly smaller percentage of female than of male illiteracy in I920. It is to be noted that the decrease in actual number of illiterates began one decade earlier for women than for men. Table V shows the actual number of illiterates and percentage of illiteracy for males and females, for the years 1870 to 1920, for the nine divisions of the United States. All show a greater percentage of female illiteracy in 1870, which slopes downward thereafter at a more rapid pace than male illiteracy, so that in only three divisions, the New England, the Middle Atlantic, and the Mountain divisions, is the percentage of female illiteracy greater than that of male illiteracy, in I920. In the New England and Middle Atlantic divisions this may be largely explained by the higher percentage of females than males in the advanced age period' and also by the higher percentage of illiteracy among foreign-born females than males.2 'Fourteenth Census of the United States, vol. II, chap. XII, Table XII. 2The relation between age and illiteracy is fully discussed in the following chapter. TABLE IV ILLITERACY IN THE UNITED STATES,* BY SEX, 1870 TO 19203 Number of Males Number of Male Per Cent of Male Number of Number of Per Cent of Year 10 Years of Age Illiterates Females 10 Years Female Female and Over of Age and Over Illiterates Illiterates 1870........ 14,258,866 2,603,888 18.3 13,970,079 3,054,256 21.9 1880........ 18,735,980 2,966,421 15.8 18,025,627 3,273,537 18.2 1890........ 24,352,659 3,008,222 1.2.4 23,060,900 3,316,480 14.4 1900........ 29,703,440 3,011,224 10. 1 28,246,384 3,168,845 11.2 1910......... 37,027,558 2,814,950 7.6 34,552,712 2,701,213 7.8 1920........ 42,289,969 2,540,209 6.0 40,449,346 2,391,696 5.9 *Figures for 1890 are exclusive of illiterate persons in Indian Territory and on Indian Reservations, areas especially enumerated but for which statistics are not available. 'Fourteenth Census, vol. II, chap. XII, Table I. H t~j TABLE V ILLITERACY IN THE DIVISIONS OF THE UNITED STATES, BY SEX, 1870 TO 19204 Number of Number of Per Cent of Number of Number of Per Cent of Division Year Males 10 Years Male Male Females 10 Femae Female of Age and Over Illiterates Illiterates Years of Age Illiterates Illiterates and Over New England............. Middle Atlantic........... East North Central........ 1870 1880 1890 1900 1910 1920 1870 1880 1890 1900 1910 1920 1870 1880 1890 1,348,135 1,559,594 1,889,307 2,228,540 2,649,897 2,940,130 3,286,841 3,967,566 5,015,259 6,108,053 7,863,584 8,890,489 3,388,148 4,302,506 5,314,088 81,668 85,791 112,626 132,411 140,326 140,280 202,924 213,640 294,560 342,866 442,488 413,458 241,272 245,319 281,016 6.1 5.5 5.9 5.9 5.3 4.8 6.2 5.4 5.9 5.6 5.6 4.7 7. 1 5.7 5.3 1,425,202 1,660,262 1,970,421 2,296,062 2,681,017 3,005,859 3,370,614 4,082,668 5,013,390 6,059,506 7,582,931 8,775,865 3,198,235 4,036,669 5,003,695 114,295 112,715 130,778 139,991 140,480 149,420 313,390 287,223 322,025 361,268 431,324 451,924 301,176 261,967 307,949 8.0 6.8 6.6 6.1 5.2 5.0 9.3 7.0 6.4 6.0 5.7 5. 1 9.4 6.5 6.2 tTI tTI H tri 0 H 0T1 CI, tC q M C3 H ttj ttf M Cn West North Central....... South Atlantic........... East South Central........ 1900 1910 1920 1870 1880 1890 1900 1910 1920 1870 1880 1890 1900 1910 1920 1870 1880 1890 1900 1910 1920 6,387,365 7,529,768 8,837,101 1,452,486 2,381,598 3,513,995 4,140,550 4,807,164 5,112,443 2,030,149 2,588,035 3,178,769 3,798,278 4,528,942 5,282,930 1,532,085 -1,900,639 2,313,978 2,758,148 3,116,286 3,348,984 259,427 262,137 262,638 154,013 170,253 176,777 153,176 138,030 101,744 904,627 1,003,565 929,096 879,065 723,570 637,980 653,063 746,439 674,991 665,392 542,291 447,071 4.1 3.5 3.0 10.6 7.2 5.0 3.7 2.9 2.0 44.6 38.8 29.1 23. 1 16.0 12. 1 42.7 39.3 29.2 24.1 17.4 13. 3 6,055,937 7,039,181 8,293,685 1,253,565 2,040,067 3,077,835 3,698,014 4,290,147 4,777,297 2,177,249 2,698,610 3,237,152 3,817,881 4,483,884 5,230,517 1,576,931 1,930,462 2,294,257 2,716,079 3,062,292 3,328,245 274,872 229,713 232,832 169,456 175,481 198,526 170,847 125,108 91,477 1,038,539 1,126,265 1,055,792 942,281 720,724 574,962 740,132 819,323 758,678 699,543 529,809 398,388 4.5 3.3 2.8 13.5 8.6 6.5 4.6 2.9 1.9 47.7 41.7 32.6 24.7 16.1 11.0 46.9 42.4 33.1 25.8 17.3 12.0 cr4 tt" M 0 TABLE V-Continued Number of Number of Per Cent of Number of Per Cent of Division Year Males 10 Years Male Male Females 10 Female of Age and Over Illiterates Illiterates YearspfAge Illiterates Illiterates and Over West South Central........ Mountain............... Pacific.................. 1870 1880 1890 1900 1910 1920 1870 1880 1890 1900 1910 1920 1870 1880 1890 733,720 1,169,117 1,663,636 2,418,607 3,334,078 3,999,088 152,090 324,638 548,004 737,787 1,185,047 1,398,659 335,212 542,286 915,623 307,354 414,185 423,764* 467,241 424,354 400,795 37,172 44,728 51,445 60,413 75,242 66,395 21,795 42,501 66,947 41.9 35.4 25.5 19.3 12.7 10.0 24.4 13.8 9.4 8.2 6.4 4.7 6.5 7.8 7.3 705,484 1,076,025 1,527,616 2,231,381 3,059,965 3,740,448 85,548 175,803 342,248 538,289 869,202 1,165,804 177,251 325,061 594,286 323,846 422,642 461,438 486,403 421,250 372,842 37,767 45,680 50,458 62,488 65,495 66,264 15,655 22,241 30,836 45.9 39.3 30.1 21.8 13.8 10.0 44.2 26.0 14.7 11.6 7.5 5.7 8.8 6.8 5.2 tTI t-1 Ctl H C', M tii M Cn 1900 1,126,112 1910 2,012,792 1920 2,480,145 51,233 66,512 69,848 4.6 833,235 3.3 1,484,093 2.8 2,131,626 31,152 37,310 53,587 3.7 2.5 2.5 *Figures for 1890 are exclusive of illiterate persons in Indian Territory and on Indian Reservations, areas especially enumerated, but for which statistics are not available. 4Fourteenth Census, vol. II, chap. XII, Table VII; Compendium of the Eleventh Census, Part III, pp. 301,302; Ninth Census, vol. I, Tables IX, XXVI (data computed). C —q II t-l 04 z M ~d - C/) ~tJ x >4 TABLE VI ILLITERACY IN THE UNITED STATES, BY SEX, BY STATES, FOR 19206 Number of Number of Per Cent of Number of Per Cent of State Males 10 Years Male Male Femal e Female of Age and Over Illiterates Illiterates Years of Age Illiterates Illiterates and Over Maine............... 314,575 12,421 4.0 306,658 7,819 2.5 New Hampshire......... 181,286 8,246 4.6 180,644 7,542 4.2 Vermont............... 144,525 5,156 3.6 139,947 3,332 2.5 Massachusetts.......... 1,514,904 68,423 4. 5 1,591,865 78,184 4.9 Rhode Island........... 237,116 14,168 6.0 246,672 17,144 7.0 Connecticut............ 547,724 31,866 5.8 540,073 35,399 6.6 New York.............. 4,186,818 188,353 4.5 4,215,968 236,669 5.6 New Jersey............. 1,256,332 61,546 4.9 1,237,914 66,115 5.3 Pennsylvania........... 3,447,339 163,559 4. 7 3,321,983 149,140 4. 5 Ohio................... 2,382,040 72,627 3.0 2,242,416 58,379 2.6 Indiana................ 1,198,722 28,864 2.4 1,157,492 23,170 2.0 Illinois................. 2,647,505 86,698 3.3 2,537,438 87,289 3.4 Michigan............ 1,536,629 48,173 3.1 1,358,977 39,873 2.9 Wisconsin............. 1,072,205 26,276 2.5 997,362 24,121 2.4 ~tI r) C) H z H H ci M t:: M co Minnesota.............. 986,877 17,413 1.8 890,255 17,074 1.9 Iowa................... 980,360 11,353 1.2 932,795 9,327 1.0 Missouri............... 1,385,747 45,444 3.3 1,352,024 37,959 2.8 North Dakota.......... 251,989 4,681 1.9 218,221 5,256 2.4 South Dakota........... 258,683 3,806 1.5 223,512 4,303 1.9 Nebraska............... 528,290 6,999 1.3 484,262 6,785 1.4 Kansas................. 720,497 12,048 1.7 676,228 10,773 1.6 Delaware............... 91,802 5,697 6. 2 87,128 4,811 5. 5 t Maryland.............. 582,933 33,435 5.7 576,020 30,999 5.4 Virginia............... 886,493 107,374 12.1 862,375 87,785 10.2 West Virginia........... 570,617 40,896 7.2 512,778 28,517 5. 6 North Carolina.......... 917,883 125,302 13.7 926,790 116,301 12.6 South Carolina.......... 604,224 110,425 18.3 615,092 110,242 17.9 Georgia.............. 1,069,254 173,254 16.2 1,080,976 155,584 14.4 Florida................. 386,150 37,252 9.6 365,637 34,559 9.5 Kentucky.............. 933,175 86,495 9.3 904,259 68,519 7.6 Tennessee............. 885,962 98,852 11.2 884,810 83,777 9.5 Alabama............... 861,344 140,991 16.4 869,077 137,091 15.8 Mississippi.............. 668,513 120,733 18. 1 670,099 109,001 16.3 Arkansas............... 667,972 63,959 9.6 634,933 57,878 9. 1 Louisiana.............. 684,958 148,081 21.6 681,108 151,011 22.2 Oklahoma.............. 797,753 32,347 4.1 716,198 24,517 3.4 Texas................. 1,848,405 156,408 8.5 1,708,209 139,436 8.2 TABLE VI-Continued I State Number of Males 10 Years of Age and Over Number of Male Illiterates Per Cent of Male Illiterates Number of Females 10 Years of Age and Over Number of Female Illiterates Per Cent of Female Illiterates ~~ Montana............... 235,586 5,357 2.3 185,857 4,187 2.3 Idaho.................. 179,948 3,085 1.7 146,103 1,839 1.3 Wyoming............... 88,316 2,180 2.5 62,677 969 1.5 Colorado............... 395,632 11,587 2.9 351,853 12,621 3.6 New Mexico............ 143,826 18,235 12.7 123,769 23,402 18.9 Arizona............... 143,651 19,984 13.9 111,810 19,147 17.1 Utah.................. 172,295 3,678 2.1 159,235 2,586 1.6 Nevada................ 39,405 2,289 5.8 24,500 1,513 6.2 Washington............. 605,288 10,479 1.7 496,641 8,047 1.6 Oregon................. 343,059 5,589 1.6 295,928 3,728 1.3 California.............. 1,531,798 53,780 3.5 1,339,057 41,812 3.1.a ON I: t-4 c, 1 -q V)!3 C/) M 02 tv~ 5Fourteenth Census of the United States, vol. II, chap. XII, Table XII. ILLITERACY AND SEX 37 In the Mountain Division, the southwestern states with their high percentages of female illiterate foreign-born whites, Indians, and descendants of Mexicans raise the female rate above the male for the division as a whole.6 In the West South Central Division, the virtual equality of the illiteracy rate is counterbalanced by the more downward slope of the trend line of the female illiteracy. The factor of nationality entering here is discussed in the chapter on race and nationality. Upon further analysis of the sex difference in illiteracy, by states, it is observed that thirty-four of the forty-eight states have a higher percentage of male illiteracy than of female illiteracy in 1920 (see Table VI). Of the remaining fourteen states, three have a difference of I/io of I per cent. The fourteen states having a higher female illiteracy rate arrange themselves into four groups, group one containing Massachusetts, Rhode Island, Connecticut, New York, New Jersey, and Illinois. As has been stated in connection with the divisions, these highly industrialized areas contain a large percentage of foreign-born women, among whom the percentage of illiteracy is high, as well as a predominance of females among the aged. Group two contains the states of Minnesota, North Dakota, South Dakota, and Nebraska, states with low illiteracy rates, with a relatively large ratio of northwestern European immigrants or their immediate descendants. The females of these foreign-born have high illiteracy rates, particularly in old age.7 In group three, Louisiana is the only state listed. The. higher female illiteracy rate is largely a function of the older Negro women and the older native white women of native parentage as well.8 In group four, Colorado, New 6Ibid., Table XII. 7bid. 8lbid. 4 TABLE VII ILLITERACY IN THE UNITED STATES AND ITS DIVISIONS, BY SEX, BY URBAN AND RURAL ENVIRONMENT, FOR 19209 Males Females Section of United States Number 10 Number 10 Number Per Cent Number Per Cent Years of Ag Illiterate Illiterate Years of Age Illiterate Illiterate and Over and Over United States Urban......... 22,009,152 926,289 4.2 21,969,424 1,028,823 4. 7 Rural.......... 20,280,817 1,613,920 8.0 18,479,922 1,362,873 7.4 tC)t 01 On a) M t,~-q M New England Urban......... Rural.......... Middle Atlantic......... Urban......... Rural.......... East North Central Urban........ Rural.......... 2,307,385 632,745 6,636,690 2,253,799 114,664 25,616 316,673 96,785 5.0 4.0 4.8 4.3 3.3 2.4 2,405,753 600,106 6,683,251 2,092,614 130,635 18,785 374,553 77,371 5.4 3. 1 5.6 3.7 3.3 1.9 5,386,447 3,450,654 179,490 83,148 5,157,081 3,136,604 172,138 60,694 West North Central Urban......... 1,941,553 39,008 2.0 1,961,536 40,407 2.1 Rural.......... 3,170,890 62,736 2.0 2,815,761 51,070 1.8 South Atlantic Urban......... 1,712,684 103,868 6.1 1,813,746 123,187 6.8 Rural.......... 3,570,246 534,112 15.0 3,416,771 4'1,775 13.2 East South Central t Urban......... 784,007 59,146 7.5 848,076 67,208 7.9 Rural.......... 2,564,977 387,925 15.1 2,480,169 331,180 13. 4 West South Central Urban......... 1,213,925 68,524 5.6 1,201,480 78,296 6. 5 Rural.......... 2,785,163 332,271 11.9 2,538,968 294,546 11.6 Z Mountain Urban......... 507,654 12,565 2.5 476,482 12,830 2.7 7 Rural.......... 891,005 53,830 6.0 689,322 53,434 7.8 Pacific, Urban......... 1,518,807 32,351 2.1 1,422,019 29,569 2.1 Rural.......... 961,338 37,497 3.9 709,607 24,018 3.4 9Ibid., Table XXII. 40 ILLITERACY IN THE UNITED STATES Mexico, Arizona, and Nevada are placed. The explanation for the illiteracy rate in these states has already been given in connection with the Mountain Division. In comparing sex illiteracy by urban and rural environments for the United States as a whole, one notes (Table 7) the fact that the percentage of illiteracy is greater for females in the urban environment, while the reverse is true in the rural environment. This once more resolves itself into the relatively higher percentage of foreign-born female illiteracy, for the foreign-born largely congregate in the industrialized, urban centers, and into the higher percentage of aged women of all classes in urban areas, in general, as contrasted with rural areas. Analyzing these differences by divisions, the greater female percentage of illiteracy in the urbanized areas is sound in every division except that of the Pacific Coast ftates, where the difference of 5/Io of I per cent is quite small. Contrariwise, in the rural areas, the ratio of female illiteracy is smaller in every division except that of the Mountain states where the much larger percentage of illiteracy on the part of the females has already been discussed. Thus the apparent difference in illiteracy rates by sex largely resolves itself into the sex distribution according to urban and rural environments, age distributions, race and nationality distributions, and lastly the institutionalized educational facilities for the sexes both in the United States and in Europe. When these factors are adequately taken into consideration, sex differences in regard to illiteracy practically disappear. CHAPTER IV ILLITERACY AND AGE With the general extension of the public school system in the United States, one would expect a diminishing percentage of illiteracy as one proceeds from the older age groups to the younger. Table VIII supports this conclusion for both sexes and for each sex taken separately. A comparison of the sexes in regard to age and illiteracy in the same table bears out the statement of the preceding chapter, namely, that several decades ago the educational opportunity for women was less than for men. The comparatively greater grasp of the opportunity by female children than males in recent years is also to be noted. There is a greater percentage of male illiteracy until the 25-34 year period.' From then on the percentage of illiteracy for males is less than for females. The data for age are subject to certain qualifications. There are concentrations of numbers around five-year periods, and to a certain extent ages are reported as of even years, i.e., 12, 14, etc. It would appear that adult women tend to understate their ages and that there is a slight tendency on the part of young men to bring their ages up to twenty-one. Among the very aged there is a tendency to overstate the actual age. The census reports, as utilized in this study, largely negate these weaknesses. In general, the degree of inaccuracy is greater for adults than for youths and children. The broad age classifications, utilized in this study, are for five-year periods only up to, 'Census illiteracy data for five-year periods are available only up to the twenty-fifth year of age. TABLE VIII ILLITERACY IN THE UNITED STATES, BY AGE PERIODS, BY SEX, FOR L 19202 - - 11 Both Sexes I Males Age Period 10-14 Years...... 15-19 Years...... 20-24 Years..... 25-34 Years...... 35-44 Years...... 45-54 Years...... 55-64 Years...... 65 and Over...... Age Unknown..... Total Population 10,641,137 9,430,556 9,277,021 17,157,684 14,120,838 10,498,493 6,531,672 4,933,215 148,699 Number Illiterate 246,360 283,316 392,853 961,200 988,961 857,776 594,573 591,385 15,481 Per Cent Illiterate 2.3 3.0 4. 2 5.6 7.0 8.2 9. 1 12.0 10.4 Total Population 5,369,306 4,673,792 4,527,045 8,669,016 7,359,904 5,653,095 3,461,865 2,483,071 92,875 Number Illiterate 141,576 171,489 203,773 486,217 509,107 453,950 292,511 273,000 8,586 Per Cent Illiterate 2.6 3.7 4. 5 5.6 6.9 8.0 8.4 11.0 9. 2 Females Total Number Per Cent Population Illiterate Illiterate 5,271,831 104,784 2.0 4,756,764 111,827 2.4 4,749,976 189,080 4.0 8,488,668 474,983 5.6 6,760,934 479,854 7.1 4,845,398 403,826 8.3 3,069,807 302,062 9.8 2,450,144 318,385 13.0 55,824 6,895 12.4 t-4 1-4 tTc Cl C-q M M W) 2Fourteenth Census of the United States, vol. II, computed from chap. III, Table I, and chap. XII, Table XII. ILLITERACY AND AGE 43 but not including, twenty-five years, the years of greatest accuracy. From the twenty-fifth year up to, but not including, the sixty-fifth, the age groups are by ten-year periods. While this makes impossible certain comparisons on a five-year basis, it controls the tendency to concentrate around the ages of 30, 40, 50, and 6o years. The aged classification is "65 years and over," controlling the errors in age as reported after sixty-five. Moreover, the broad generalizations in regard to age are believed to be justifiable. An analysis of Table IX further reveals, by divisions, the differential illiteracy according to age, with a gradual increase in the percentage of illiteracy as one approaches the age period of sixty-five years and over. The precipitous rise in percentage by the New England, Middle Atlantic, and East North Central Divisions after the two earliest age periods is due to the fact that immigrants are only in small part children, and the young adult (as well as the child) immigrant is more literate than his older immigrantcountrymen.3 The data in Table X may be utilized as a partial reflection of the educational opportunities during the decades preceding the taking of the Fourteenth Census. Native whites of native parentage are compared here, thus eliminating the colored and foreign-born population. An interpretation of the I920 data brings out the fact that, of those persons who were born in i855 or prior to that year, 6.2 per cent of the survivors are illiterate. Of those born during the years i906-i9io, only i.i per cent of the survivors are illiterate. 3See Fourteenth Census of the United States, vol. II, chap. XII, Table XII, for statistical proof of this. TABLE IX PERCENTAGE OF ILLITERACY IN THE DIVISIONS OF THE UNITED STATES, BY AGE, BY SEX, FOR 19204 Both Sexes Ae Period New Middle East North WestNorth South East South West South Mou tain f Age Period_ Mountain Pacific England Atlantic Central Central Atlantic Central Central 10-14 Years..... 0.3 0.3 0.2 0.4 4.9 6.2 5.9 2.2 0.4 15-24 Years..... 2.0 1.6 0.9 0.7 8.1 8.1 7.4 3.8 1.5 25-34 Years.... 6.4 6.0 3.1 1.5 9.5 10.4 8.6 4.8 2.9 35-44 Years..... 6.9 7.3 3.8 2.1 12.6 13.5 10.8 5.7 3.3 45-54 Years.... 6.3 6.9 3.9 2.7 17.1 19.4 14.4 6.9 3.2 55-64 Years.... 6.2 6.8 4.5 3.7 21.1 23.9 17.1 7.7 3.4 65 and Over..... 6.5 7.2 6.7 6. 6 27.6 30. 4 22.4 11.1 4.6 Males H t_1 C-) H z Pd n H H 'Co M c3 3 m ttj Cn 10-14 Years..... 15-24 Years..... 25-34 Years..... 35-44 Years..... 45-54 Years..... 0.3 1.7 5.9 7.0 6.5 0.3 1.4 5.5 7.2 6.8 0.3 0.9 3.1 4.2 4.2 0.4 0.9 1.6 2.3 2.7 5.8 9.9 10.2 12.4 16.3 7.2 10.1 11.4 13.4 18.7 6.5 8.3 8.7 10.2 13. 1 2.2 3.7 4.4 5.1 6.1 0.4 1.6 3.0 3.4 3.4 55-64 Years..... 6.0 6.2 4.4 3.4 19.6 22.0 15.1 6.3 3.5 65 and Over..... 6.3 6. 3 6.0 5.5 26.0 27.9 20.0 9.6 4.3 Females 10-14 Years..... 0.2 0.2 0.2 0.3 4.0 5.1 5.4 2.2 0.4 15-24 Years..... 2.2 1.9 0.9 0.6 6. 4 6. 2 6.5 3.9 1.3 25-34 Years..... 6.8 6.4 3.0 1.3 8.8 9.4 8.4 5.3 2.7 35-44 Years..... 6.7 7.4 3.5 1.9 12.7 13.5 11.4 6.5 3.1 45-54 Years..... 6.0 7.0 3.6 2.7 17. 9 20.2 16. 0 8.1 3.0 55-64 Years..... 6.3 7.5 4.7 4.1 22.9 26.0 19.7 9.7 3.3 65 and Over..... 6. 6 8.0 7.5 7. 8 29.2 33.1 25.3 13.1 4.9 4Fourteenth Census of the United States, vol. II, chap. XII, Table XII. c-q I.1-(q H C) M tv MI 46 ILLITERACY IN THE UNITED STATES Of those native white of native parentage born in 1855 or before, born in 1856-1865 born in 1866-1875 born in I876-I885 born in I886-I895 born in I896-I905 born in I906-I9IO 6.2% 4.9% 3.7% 2.6% 1.9% i.6% I.I 0 were illiterate in 1920 were illiterate in 1920 were illiterate in I920 were illiterate in I920 were illiterate in I920 were illiterate in 1920 were illiterate in 1920 Thus were no further efforts made to reduce the illiteracy rate, and provided conditions remained the same, it may be assumed that the illiteracy rate for native whites of native parentage would approach I per cent, as those persons in the older age periods died. Table X also presents interesting aspects when the illiteracy rate of one age group for one census year is compared TABLE X PERCENTAGE OF ILLITERACY IN THE UNITED STATES, BY AGE, FOR NATIVE WHITES OF NATIVE PARENTAGE, 1890 TO 1920*5 Age 1890 1900 1910 1920 10-14 Years........ 6. 7 4.4 2.2 1.1 15-24 Years........ 5.7 4.3 2.6 1.6 25-34 Years........ 6.9 4.6 3.0 1.9 35-44 Years........ 8. 1 6.4 3.8 2.6 45-54 Years........ 8.4 8.0 5.6 3.7 55-64 Years........ 10. 2 8.0 6.7 4.9 65 Years and Over... 12. 7 10. 7 7. 6 6. 2 *Data in the corresponding age groups are not available for 1870 and 1880. with the next highest age group of the succeeding census. For example, in I890, the age period 15-24 years had an illiteracy rate of 5.7 per cent. In 900o, the next age period, 5Ibid., Tables III and XII; Thirteenth Census, vol. I, chap. XIII, Tables VI, VIII and XXVII; chap. IV, Table XVI. ILLITERACY AND AGE 47 25-34 years, composed of the survivors of the 15-24 year period, was only 4.6 per cent illiterate. Thus there was a decrease in rate of I.I per cent. In I9I0 the 35-44 year period had an illiteracy rate of 3.8 per cent, a decrease of 0.8 per cent. In I920 the survivors of the original group, now 45-54 years of age, were 3.7 per cent illiterate. Starting with other age periods in I890, the same tendency to decrease is apparent. Is this due to schools for adults or to the fact that the necessities of adjustment to modern culture tend to result in the escape from illiteracy of a significantly high percentage in the course of a thirty or forty year period? The latter explanation appears more justifiable due to the late introduction of organized efforts to reduce the illiteracy rate in the various states. Thus starting with the indicated age period in I890, each group became less illiterate with each succeeding census year.6 In conclusion, the statistics of illiteracy and age would seem to show that illiteracy is decreasing not only because of increased school facilities, but also as a result of the mortality of the aged. As has been shown, it is the aged groups who have the highest rates of illiteracy. 6The single apparent exception is the 10.2 per cent for the 55-64 year period in 1890 and the 10.7 per cent for 65 years and over in 1900. This latter rate is not comparable with the preceding rate due to the fact that it includes not only those 65-74 years of age but also those 75 years of age and over. CHAPTER V ILLITERACY AND URBAN AND RURAL ENVIRONMENT The illiteracy rates differ so greatly between urban and rural communities that a separate discussion of the factors is entered into at this point. By "urban" and "rural," one may proceed only by the not always adequate census dichotomy. The United States Census Bureau classifies as urban "all incorporated places (and all towns in Massachusetts, Rhode Island, and New Hampshire) having 2,500 inhabitants or more."' The remainder of the country is treated as rural. Obviously this is merely a working basis as there is actually no clear line of demarcation. The statitical line has been drawn, however, after careful consideration by the Bureau of the Census, and is adequate for the needs of the present study. An urban illiteracy rate of 4.4 per cent for the United States as a whole in 1920 is to be compared with a rural illiteracy rate of 7.7 per cent. The rural illiteracy rate of 7.7 per cent for 1920 compares favorably with the rural illiteracy rate of IO.I per cent for I9I0 as does the urban rate of 4.4 per cent with that of 5.I per cent for the preceding census year. Illiteracy data according to urban and rural environment are not available for 900o or preceding decades. In an analysis of the available data according to the nine divisions, the same results are found. In each section the decrease in illiteracy is greater for rural than for urban 'Fourteenth Census of the United States, vol. II, chap. I, p. 20. ILLITERACY AND ENVIRONMENT 49 TABLE XI ILLITERACY IN THE UNITED STATES, BY URBAN AND RURAL ENVIRONMENT, FOR 1910 AND 1920*2 Urban Year Year Total Population Number Per Cent 10 Years of Age Illiterate Illiterate and Over 1910 34,278,790 1,748,830 5.1 1920 43,978,576 1,955,112 4.4 Rural Year Total Population Number Per Cent 10 Years of Age Illiterate Illiterate and Over 1910 37,301,480 3,767,333 10.1 1920 38,760,739 2,976,793 7.7 *Earlier census reports do not give illiteracy according to urban and rural environment. regions, comparing the I920 with the I9IO results. There is an actual increase of o.I per cent of illiteracy in the urban rate for the Pacific division. Further analysis reveals that, of the three states in this division, the urban illiteracy rate of 2.4 per cent has remained the same for California from I9I0 to I920; for Oregon, the urban illiteracy rate has increased from 1.3 per cent in 19Io to 1.5 per cent in I920; for Washington, the urban illiteracy rate has increased from 1.3 per cent in I910 to 1.5 per cent in 1920. The apparent reversal of trend is traceable to the increase in the percentage of illiteracy of the foreign-born white in 2lbid., chap. XII, Table XXI. TABLE XII ILLITERACY IN THE DIVISIONS OF THE UNITED STATES, BY URBAN AND RURAL ENVIRONMENT, FOR 1920 AND 19103 1920 1910 r Division Number 10 Number Number 10 umber Per Cent Number Per, Cent Number Per Cent Years of Age Illiterate Illiterate Illiterate and Over and Over New England Urban......... 4,713,138 245,299 5.2 4,064,027 227,841 5.6 H Rural.......... 1,232,851 44,401 3.6 1,266,887 52,965 4.2 Middle Atlantic c Urban......... 13,319,941 691,226 5.2 11,033,550 644,618 5. 8 Rural.......... 4,346,413 174,156 4.0 4,412,965 229,194 5.2 t East North Central C. Urban......... 10,543,528 351,628 3.3 7,831,590 277,444 3.5 H Rural...........6,587,258 143,842 2.2 6,737,359 214,406 3.2 -3 M CI) West North Central Urban........... 3,903,089 79,415 2.0 3,203,714 86,958 2.7 Rural.......... 5,986,651 113,806 1.9 5,893,597 176,180 3.0 /....................... South Atlantic Urban......... Rural.......... East South Central Urban........ Rural......... West South Central Urban......... Rural.......... Mountain Urban........ Rural.......... Pacific Urban......... Rural.......... 'Ibid., Table XXI. 3,526,430 6,987,017 1,632,083 5,045,146 2,415,405 5,324,131 984,136 1,580,327 2,940,826 1,670,945 227,055 985,887 126,354 719,105 146,820 626,817 25,395 107,264 61,920 61,515 6.4 14.1 7.7 14.3 6.1 11.8 2.6 6.8 2. 1 3.7 2,493,359 6,519,467 1,279,677 4,898,901 1,562,545 4,831,498 772,572 1,281,677 2,037,756 1,459,129 211,760 1,232,534 122,477 949,623 112,889 732,715 23,962 116,775 40,881 62,941 8.5 18.9 9.6 19.4 7.2 15.2 3. 1 9. 1 2.0 4.3 -C C4 t-1 til -t z M,-]! Ptx 1-3 U-i I.^ 52 ILLITERACY IN THE UNITED STATES urban areas in all three Pacific division states.4 The California urban illiteracy rate for foreign-born whites increased from 7.1 per cent in I9IO to 8.0 per cent in I920; that for Oregon increased from 4.2 per cent in I91O to 5.3 per cent in I920; the rate for Washington increased from 4.0 per cent in I9IO to 4.5 per cent in I920. In the states of the New England, Middle Atlantic, East North Central, and West North Central Divisions (see Table XIII), the percentage of illiteracy in I920 is higher for the urban than for the rural communities. In all four divisions this is traceable to the large proportion of illiterate foreign-born whites in the urban population.5 Missouri, North Dakota, and South Dakota are the exceptions, all three having higher rural illiteracy rates due to the normally greater opportunity for education in urban communities. In Missouri and South Dakota, the higher foreign-born white urban illiteracy rates are more than offset by the small ratio of foreign-born whites to the population of these states. Comparison of the urban and rural illiteracy rates for states in the remaining five divisions shows that all the states have larger rural illiteracy rates than urban illiteracy rates, with the exception of Oregon, where the difference is i/io of I per cent. The relatively larger foreign-born white illiteracy rate in urban than in rural areas for Oregon (5.3 per cent as compared with 5.0 per cent) is the important factor in this slight difference. The results from the dichotomous division into urban and rural environment are further tested by a four-part division into (a) rural districts, (b) cities of from 2,500 to 25,000 inhabitants, (c) cities of from 25,000 to IO00000 4See Thirteenth Census of the United States, vol. I, chap. XIII, Table XXIII, and Fourteenth Census of the United States, vol. II, chap. XII, Table XXII. 5Ibid. ILLITERACY AND ENVIRONMENT 53 TABLE XIII PERCENTAGE OF ILLITERACY IN THE UNITED STATES, BY STATES, BY URBAN AND RURAL ENVIRONMENT, 19206, j \ State Urban Rural State Urban Rural I'- 1 — 11 [II I Maine................ New Hampshire....... Vermont............. Massachusetts........ Rhode Island......... Connecticut........... New York............ New Jersey........... Pennsylvania........ Ohio................. Indiana.............. Illinois............... Michigan............. Wisconsin............ Minnesota............ Iowa................. Missouri.............. North Dakota......... South Dakota......... Nebraska............. Kansas............... Delaware............. Maryland............ Virginia............. West Virginia......... North Carolina........ South Carolina........ Georgia............. Florida............... 3.5 5.3 3.9 4.8 6.5 6.6 5.5 5.3 4.6 3.2 2.6 3.9 3.4 2.6 1.9 1.4 2.6 1.5 1. 1 2.0 2.0 5.7 4.4 7.1 3.2 9.3 10.3 9.5 5.4 3. 1 2.8 2.6 3.0 5.7 5.3 2.9 4.6 4.6 2.2 1.8 2.2 2.5 2.3 1.8 0.9 3.5 2.2 1.8 1.0 1.4 6.1 7.4 13.0 7.6 14.1 20.0 17.5 12.2 Kentucky........... Tennessee........... Alabama............ Mississippi.......... Arkansas............ Louisiana............ Oklahoma........... Texas............... Montana............ Idaho............... Wyoming............ Colorado............ New Mexico......... Arizona............. Utah................ Nevada............. Washington......... Oregon.............. California........... 5.1 7.0 10.4 11.3 4.9 9.1 1.9 6.5 1.6 1.4 2.1 2.3 7.1 6.5 1.3 1.8 1.5 1.5 2.4 9.8 11.6 17.8 18.2 10.3 29.6 4.5 9.3 2.6 1.5 2.1 4.2 17.6 20.4 2.5 7.0 2.0 1.4 5.4 'Ibid., Table XXII. 5 TABLE XIV PERCENTAGES OF ILLITERACY IN THE UNITED STATES, FOR NATIVE WHITES OF NATIVE PARENTAGE, BY STATES, FOR PRINCIPAL URBAN AND RURAL DIVISIONS, 19207 Places of Places of State Rural Districts 2,500to 25,000 to 100,000 25,000 100,000 and Over State Rural Districts 2,500 to 25,000 to 25,000 100,000 100,000 and Over Maine............... New Hampshire....... Vermont............. Massachusetts........ Rhode Island......... Connecticut......... New York............ New Jersey........... Pennsylvania......... Ohio................. Indiana.............. Illinois............... Michigan............. 1.6 0.6 1.1 0.5 2.2 0.8 1.0 1.5 1.3 1.4 1.6 1.5 0.8 0.5 0.7 0.9 0.4 0.5 0. 3 0.6 0.6 0.6 0.9 1.2 1.2 0.9 0.4 0.2 0. 2 0.5 0.1 0. 5 0.2 0. 0.7 1.3 0. 5 0.4 0.2 0.2 0.1 0.2 0.3 0.3 0.5 0.8 0.1 0.2 Kentucky........... Tennessee........... Alabama........... Mississippi......... Arkansas........... Louisiana.......... Oklahoma........... Texas............... Montana............ Idaho............... Wyoming............ Colorado............ New Mexico......... 8. 8 8. 8 7.7 4.0 5.2 16.5 3.0 2.7 0.4 0.3 0.4 2.5 13.2 3. 1 3.7 3.3 1.6 1.5 4.9 1. 1 1.4 0.2 0.3 0.3 1.3 6.4 1.6 3.9 0.6 2. 1 0.4 0.4 0. 7 0.2 0.9........ 1.3 0.7 0.9 0.8 0.3........ t-1 t-1 C M Pq M tt M Wisconsin........... 0.7 0.3 1 0.3 1 0.1 Minnesota............ Iow a................. Missouri.............. North Dakota......... South Dakota......... Nebraska............. Kansas............... Delaware............. Maryland............ Virginia.............. West Virginia......... North Carolina........ South Carolina........ Georgia.............. Florida............... 0.3 0. 5 3.0 0.3 0.4 0.5 0.6 3.2 3.3 7.6 5.8 9.2 7. 7 6. 7 4. 3 1.1 0.7 1.6 0.3 0.2 0. 3 0.6 2.1 1.9 3.9 1.5 4.5 2.9 2.9 0.5 0.2 0.4 1.6 0.1 0.3 0. 3 1.7 1.7 2.1 2.9 2.4 2.3 0.8 0.1 0.4 0.4 0.2 0.9 0.5 0.6 1.1 1.3 a...... Arizona............. Utah................ Nevada............. Washington......... Oregon.............. California.......... 1.6 0.4 0.5 0.4 0.4 0.7 1.0 0.7 0.2 0.4 0.6 0.4 0.3 0.2 0.2 0.2........ 0.2........ 0.1 0.2 0.2 t-1 t-1 ti 0 z z. 7Ibid., computed from Tables XIV, XIX, XXII. 56 ILLITERACY IN THE UNITED STATES inhabitants, and (d) cities of IOO,OOO or more inhabitants. The data, when computed for all classes, are obscure, largely due to the incidence of colored and foreign-born. In order to eliminate these factors, the data are computed for native whites of native parentage. The general trend of a lower illiteracy rate as one goes from the rural to the quantitatively most urbanized communities is apparent (Table XIV). Within the limits of the available data, however, it is evident that the difference between urban and rural rates is diminishing. For all classes the difference in I9IO was 5 per cent, in I920 only 3.3 per cent. When race and nationality are eliminated by considering only native whites of native parentage, the difference between urban-rural rates of illiteracy in I9IO is found to be 4.5 per cent as compared with a difference of 3.0 per cent in I920.8 Part II of the present study carries the investigation of the relation between illiteracy and urbanization further.9 8For native whites of native parentage, illiteracy rates in 1910 and 1920 were as follows: Urban Rural 1910............... 0.9 5.4 1920................ 0.8 3.8 -Fourteenth Census, vol. II, chap. XII, Table XXI. 9Attention may be called here to page 40, above, which discusses the factor of age, in combination with sex and nationality, in a partial explanation of rural and urban illiteracy differences. CHAPTER VI ILLITERACY AND RACE AND NATIONALITY There are important differences in the illiteracy rates for the various race and nationality' groups. It is the purpose of this chapter to analyze these differences and to show how they contribute to the illiteracy rates of the United States as a whole and of the various geographical divisions. An analysis of the data presented in Table XV reveals the fact that taking each race and "nationality" (so far as the census differentiates) separately, all except three groups show a consistent downward trend from I870 to I920. The illiteracy rate for the foreign-born whites is actually higher in I920 than it was in I880, the year when information on this group was first available. The trend is practically horizontal. In I880 illiteracy rates were still high in all European countries while the progressive increase in the percentage of immigration from southern and eastern Europe, which lags behind western and northern Europe in educational opportunity, has seemed to keep the rate of illiteracy among the foreign-born practically the same throughout the period studied. Furthermore, the foreignborn are chiefly adults and do not readily learn to write after their arrival in America.2 The Chinese and Japanese illiteracy trends are downward over a span of three census years, but show increases 1The utilization of the terms "race" and "nationality" in this chapter necessarily follows the U. S. Bureau of the Census classifications. 2W. Talbot; Adult Illiteracy, p. 21. TABLE XV PERCENTAGE OF ILLITERACY IN THE UNITED STATES, ACCORDING TO RACE AND NATIONALITY, 1870 TO 19203 Year White Colored* Negro Native White Native White of Native Parentage Native White of For. or Mixed Parentage I I I I I I 1920.................... 1910.................... 1900.................... 1890.................... 1880.................... 1870.................... 4.0 5.0 6.2 7.7 9.4 11.5 23.0 30. 5 44.5 56.8 70.0 79.9 1 I Year 1920................... 1910................... 1900................... 1890................... 1880................... 1870................... Native White of Foreign Parentage 0.9 1.2.............. Native White of Mixed Parentage 0.7 0.9.......................................... 22.9 30.4 44. 5 57.1 81.4 Foreign Born White 13. 1 12.7 12.9 13. 1 12.0. 2.0 5.0 6.2 7.7 8.7........ 2.5 3.7 5.7 7.5........................ Indian 0.8 1.1 1.6 2.2.............. Chinese Japanese k-n 00 t-4 O3 H tl U) c3 M All Others** I I 1 34.9 45.3 56.2.............. 20.0 11.0 15.8 9.2 29.0 18.2., I............ 13.0 39.9.............................. *Persons of Negro descent, Indians, Chinese, Japanese, and "all others." **"All Others" includes Hindus, Filipinos, Koreans, Maoris. 3Fourteenth Census, vol. II, chap. XII, Tables I and IV; Thirteenth Census, vol. I, chap. XIII, Table IV. ILLITERACY AND NATIONALITY 59 in 1920, as compared with 191o. The chances of error in collecting data are so great that no satisfactory explanation can be made. For example, in I890, "it would appear that some assistant marshals committed the fault of returning as illiterate the Chinese who could not write English while they could read and write their own language."4 Undoubtedly the data for I920 are the most satisfactory. The highest illiteracy rate, from I900 to 1920, is recorded by the Indian group. The pre-literate culture of the Indians is still influencing the Indians' reactions to American culture. The older ones, who in their youth were not required to attend government schools, have very high rates. Furthermore, the great body of Indians is in the rural areas. The rate is rapidly decreasing, however, as the influence of the schools increases. The Negro group has the next highest illiteracy rate. This is due partially to the fact that the Negro is still largely rural-66% living in rural areas in I920-and partially to the fact that he lives largely in the southeastern section of the United States-where institutional education is not as important a part of the cultural complex as in other sections of the country. Two other factors, the attitude toward Negro education in parts of the South, and the very high rate of illiteracy among aged Negroes,5 are traceable to the social situation arising from slavery and reconstruction days in the South. The rate of illiteracy has decreased practically 50 per cent from 1900 to 1920, and with greatly increased provisions for Negro education throughout the United States, there should be a further important decrease in the coming decades. 4Ninth Census, vol. I, footnote to Table IX. 5Forty-nine and four tenths per cent of the Negroes 55 to 64 years were illiterate in 1920; sixty-eight and three tenths per cent of the Negroes 65 years and over were illiterate at that time. 60 ILLITERACY IN THE UNITED STATES The numerically smallest group-Hindus, Filipinos, Koreans, and Maoris, grouped by the census under the classification "All Others "-has sharply decreased its illiteracy rate. This is due chiefly to the low illiteracy rate among the immigrants from I9IO to I920.6 The Chinese, Japanese, and foreign-born whites have been discussed above. Turning to those native whites who are the children of either foreign or mixed parentage, there is noted, in addition to the steady decrease in the illiteracy rate, the interesting fact that the illiteracy rate for this group is actually much lower than that of the native whites of native parentage. Analysis of the data bearing on this subject, however, brings out the fact that only 42.0 per cent of the native whites of native parentage resided in urban communities in I920, whereas 69.2 per cent of native-born whites of foreign or mixed parentage lived in urban areas in the census year.7 In addition, the white population of the southeastern section of the United States is overwhelmingly native white of native parentage.8 This section of the United States, as has already been pointed out, is the section in which institutionalized education is least developed. Hence the heavy proportion of native whites of native parentage in this section penalizes this group as a whole. On the other hand, there is a small proportion of native whites of foreign or mixed parentage in this section 6Compare 1910 Census, vol. I, chap. XIII, Tables I and II, with 1920 Census, vol. II, chap. XII, Table I. 7Fourteenth Census, vol. II, chap. I, Table XX. 8Ninety-one per cent of the white population of the South Atlantic Division is native born of native parentage, as compared with three and seven tenths per cent for native whites of foreign or mixed parentage. Ninety-five and seven tenths per cent of the white population of the East South Central Division is native born of native parentage, as compared with one and eight tenths per cent for native born whites of foreign or mixed parentage. Ibid., Table VII. ILLITERACY AND NATIONALITY 61 so that the general illiteracy rate for this group is but slightly affected. As a check on this reasoning, one finds, when the rural populations for the two groups are held constant, the difference is brought down to 32/I00 of I per cent, which is the percentage by which the illiteracy rate of urban native white of native parentage exceeds the rate for native white of foreign or mixed parentage for the country as a whole.9 Thus the difference between the two groups resolves itself largely into a function of educational opportunity. The slight difference in favor of native whites of mixed parentage as compared with native whites of foreign parentage may reasonably be ascribed to the probable better opportunity afforded the child of parents, one of whom is native born. Table XVI is presented in further analysis of the differential illiteracy rates for the various race and "nationality" groups. The comparison of divisions for the native whites of native parentage reflects in large measure the educational opportunity afforded by the various sections. The high illiteracy rate for foreign-born whites, and for native born whites of foreign and mixed parentage in the West South Central Division, is traceable to the large numbers of illiterate Mexicans inhabiting this division, particularly the state of Texas. The Negro group, forming approximately one-tenth of the population and biologically less easily assimilated than the white foreign-born, is important. In three divisions, the Middle Atlantic, the Mountain, and the Pacific, the Negro rate of illiteracy is less than the rate of illiteracy for native whites of native parentage in the South Atlantic and East South Central Divisions, showing the compara9The urban illiteracy rate for native whites of native parentage is 0.81 per cent; for native whites of foreign or mixed parentage it is 0.49 per cent. Computed from data furnished in Fourteenth Census, vol. II, chap. XII, Table XXII. TABLE XVI PERCENTAGE OF ILLITERACY IN THE DIVISIONS OF THE UNITED STATES, ACCORDING TO RACE AND NATIONALITY, 192010. Division United States............... New England............... Middle Atlantic............. East North Central.......... West North Central......... South Atlantic.............. East South Central.......... West South Central.......... Mountain................. Pacific.................... Division United States............... New England............... Middle Atlantic............. East North Central........... 1.......... 1............................. --- - - -.. - - -... '..... All Classes 6.0 4.9 4.9 2.9 2.0 11. 5 12.7 10.0 5.2 2.7 Negro 22.9 7. 1 5.0 7.3 Native White of Native Parentage 2.5 0.6 0.7 1.0 1.1 5.4 6.6 3.9 2.4 0.4 Indian 34.9 12.9 15.5 20.7 Native White of Foreign Parentage 1- 1 --- Native White of Mixed Parentage 0.9 0.8 0.5 0.7 0.6 0.9 1.2 8.2 1.2 0.5 Chinese 20.0 22.3 23.0 20.0 0.7 0.7 0.4 0.5 0.4 1.0 1.8 4.2 0.7 0.3 Japanese 11.0 5.0 4.2 3.4 13.1 14.0 15.7 10.8 6.4 12.8 9. 1 29.9 12.7 8.6 All Others 13.0 2.0 6.2 4.6 Foreign Born White CI: z t"l H H cl) m P) CC (17 X tvl 1-3 Cn West North Central.................... 10.5 24.0 19.2 11.1 3.4 South Atlantic......................... 25. 2 35. 3 18. 8 2. 5 2.9 East South Central.................... 27. 9 57. 7 18.0.............. 10. 7 West South Central..................... 25. 3 18. 6 25. 9 6. 6 26. 3 M ountain............................. 5.3 58.1 21.1 16. 9 12.9 Pacific................................ 4.6 32.1 18.7 10.8 15.4 'Ibid., Tables XI, and XII. tl 0 '-4,q TABLE XVII PERCENTAGE OF ILLITERACY IN THE UNITED STATES, ACCORDING TO RACE AND NATIONALITY, BY SEX, 1870 TO 192011 Foreign Born Native White White Colored* Negro White Native White of Native Parentage Year Male Female Male Female Male Female Male Female Male Female Male Female Male Female 1920......... 4.0 23.4 22.5 23.5 22.3 11.7 14.8 2.2 1.8 2.7 2.3 1910........ 5.0 4.9 29.8 31.2 30.1 30.7 11.8 13.9 3.1 2.9 3.8 3.7 1900....... 6 6. 6 42.8 46.2 43.1 45.8 11.3 14.7 4.6 4.7 5.6 5.9 1890......... 7 8.3 3.7 59.9 54.4 59.8 11.3 15.2 5.8 6.6 6.9 8.0 1880......... 8. 6 10.2 67.3 72.7...... 1870......... 17 80.6 82.0.............. W Native White Native Wh ite of Foreign or of Foreign of Mixed Indian Chinese Japanese All Others** Mixed Paen Parentage Parentage ar Male Female Male Female Male FemaleMale Female Male Femal ale Female Male Female Year Male Female Male Female Male F'emale Male Female Male Female Male Female Male Female t-1 H M r) H H Ci) M tlf co 1920......... 0.8 1910......... 1.2 0.7 1.1 0.9 0.8 1.2 1.2 0.8 1.0 0.6 0.8 32.8 41.5 37.2 119.7 23. 1 30.2 8.7 16.2 13.0 12.5 8.6 14.1 39.9....... 49.2 1 15.0 1 900......... 1.7 1.6.................................................................. 1890......... 2.3 2.2..................................................................... 1880................................................................ 18 70.................................................................................................... 1870......... *Persons of Negro descent, Indians, Chinese, Japanese, and "all others." **"All Other" includes Hindus, Filipinos, Koreans, Maoris. "Fourteenth Census, vol. II, chap. XII, Tables I and II; Thirteenth Census, vol. I, chap. XIII, Table XIII. 4 C 0 > O C3 t~1 34 Irl 66 ILLITERACY IN THE UNITED STATES tive educational progress of a formerly and still highly illiterate group, when considered as a whole. In comparing the sexes of each race and "nationality" group, one observes that the sexes are about equal in illiteracy rates for the various native white groups and for the Negro group. In each case the difference in i920 is one in favor of the female. For the three Mongoloid groups, the Indian, the Chinese and the Japanese, there is a much higher illiteracy rate for females. For the Chinese and Japanese, the selectivity of the male group must be taken into consideration, while in the case of the Indians this is another reflection of the culture. The foreign-born whites also show a higher illiteracy rate for females. This is ascribable, it is believed, to the selectivity of the male group, on the members of which is put the brunt of establishing a footing in the United States. It is also true, as has been previously pointed out, that education for women has lagged behind that for men in most European countries. A comparison of the data for urban and rural environment (Table XVIII) reinforces the proof of more desirable educational opportunities in the urban communities. It is noteworthy that the foreign-born white group is the only one in which the difference in illiteracy rates between urban and rural areas is slight. This is traceable to the fact that the great majority of foreign-born whites emigrate after the age when they should have learned to write, and that, of those who were not literate (in the census definition) before emigrating, few have since learned. It is also worthy of note that the difference between the urban and rural illiteracy rates in i920 is less than the difference between the rates for these two divisions in i9io for every race and "nationality" group presented. TABLE XVIII PERCENTAGE OF ILLITERACY IN THE UNITED STATES, ACCORDING TO RACE AND NATIONALITY, IN URBAN AND RURAL ENVIRONMENTS, FOR 1910 AND 192012 Native White Native White Native White Foreign Born Urban and Rural All Classes of Native of Foreign of Mixed White Negro Parentage Parentage Parentage 1920 Urban.............. 4.4 0. 8 0. 5 0.4 13.0 13.4 Rural............... 7.7 3. 8 1.6 1.1 13. 3 28. 5 Native White of Foreign or Mixed Parentage 1910 Urban.............. 5.1 0.9 0.7 12.5 17.6 Rural............... 10.1 5.4 1.9 13.3 36.0 12Fourteenth Census, vol. II, chap. XII, Table XXII. c —r t-1 0-3 z ti ON 4 CHAPTER VII ILLITERACY AND SCHOOL SYSTEMS It has long been recognized that a high relationship exists between illiteracy and educational facilities, on the basis of empirical observation. One would expect this relationship, not only for present day illiteracy and present day school systems but also for present day illiteracy and school systems of ten, twenty, and even fifty or sixty years ago. Illiteracy tends to lag an entire generation behind school systems so that illiteracy rates of 1920 are largely the evidence of poor school facilities in the past. No single item in regard to school systems may be regarded as a completely adequate basis of comparison. Hence Leonard P. Ayres's An Index Number for State School Systems has been utilized. In his index Ayres combines the following ten factors: I. Per cent of school population attending school daily. 2. Average days attended by each child of school age. 3. Average number of days schools were kept open. 4. Per cent that high school attendance was of total attendance. 5. Per cent that boys were of girls in high schools. 6. Average annual expenditure per child attending. 7. Average annual expenditure per child of school age. 8. Average annual expenditure per teacher employed. 9. Expenditure per pupil for purposes other than teachers' salaries. Io. Expenditure per teacher for salaries. The index number is the average of the io figures corresponding to these 10 headings, after certain of them ILLITERACY AND SCHOOL SYSTEMS 69 have been so multiplied or divided by constants as to bring each into comparability with a standard of Ioo.l The years for which an index number is computed are 1890, I900, I9IO, and 1918. Thus a person ten years of age in 1890 would be included in the 3 5-44 year period by 1920. Since influxes of immigrants or Negroes might produce a considerable change in the population of a given area during this thirty-eight year period, and since the school indices are a constant factor for any group, illiteracy in 1920 for native whites of native parentage alone was correlated with the indices of the various periods. The following correlations2 were secured: r= -.75, for I890 r= -.74, for 1900 r= -.75, for I910 r= -.72, for I918 The four coefficients of correlation are not only quite high but they show only slight variations, the one for 1918 being slightly lower than the others. What has actually happened is that school systems have tended to develop in a relatively consistent manner so that the relationship with illiteracy for I920 has remained about the same while the indices have risen over the period involved. The scatter of the indices has shown some variation, however, the greatest diversity being noted in I918. Sy= 5.I, for I890 S = 2.7, for I900 SY= 2.2, for I9I0 Sy=8.5, for I918 'An Index Number for State School Systems, pp. 14, 15. 2The corresponding errors of the coefficients of correlation (see Appendix A) are.06 for 1890,.07 for 1900,.06 for 1910, and.07 for 1918. 6 TABLE XIX PERCENTAGE OF ILLITERACY FOR NATIVE WHITE OF NATIVE PARENTAGE IN 1920 AND INDEX NUMBERS FOR SCHOOL SYSTEMS FOR 1890, 1900, 1910, 1918, BY STATES3 Per Cent Per Cent state Illiterate, 1890 1900 191o 1918 State Illiterate 1890 1900 1910 1918State Illiate 1890 1900 1910 1918 1 ___1920 192.... North Dakota....... Washington.......... Massachusetts...... Idaho............... South Dakota........ Utah................ Montana............ Wyoming........... Minnesota........... California.......... Oregon.............. Nebraska............ Connecticut.......... Nevada............. Wisconsin.......... Rhode Island........ Iowa................ 0.26 0.29 0.30 0.31* 0.31 0.34 0.34 0.35 0.36 0.39 0.39 0.40 0.41 0.42 0.47 0.51 0.52 25.48 30.80 45.86 22.81 26.06 28.64 36.34 36.27 29.45 43.75 27.91 26.43 34.83 37.14 49.52 29.25 33.99 37.51 39. 51 31.91 35.41 43. 80 32.04 36.11 42.48 61.21 56.32 44.57 42.57 50.92 53.50 42.59 44.51 60.44 47.81 43.99 49.31 56.01 43.23 50. 84 41.45 59.17 63. 67 61.04 58.57 55.03 61.39 75.79 56.71 58.43 71.21 57.81 57. 14 59.77 59.05 51.34 56.33 61.85 Vermont........... Illinois.............. M aine.............. Arizona............ Indiana............. Colorado............ Maryland........... Delaware............ Texas............... M issouri............ Oklahoma........... Florida.............. Mississippi.......... Arkansas............ West Virginia........ Georgia............. Virginia............. 1.05 1.06 1.27 1.28 1.36 1.66 1.99 2.05 2.17 2.20 2.38 3.13 3.59 4.57 4.83 5.48 6.05 30.22 31.87 29.88 32.75 29.82 37. 83 33.30 29.30 23.23 21.88 28. 52 21.88 20.07 21.82 15.73 22.25 35.44 37.18 33.70 30. 17 36.33 41.59 35.49 30.10 24.43 20.89 23.27 22.45 20.89 20. 99 27.07 21.54 21.69 42.11 49.86 39.68 45.54 45.95 49.23 38.47 38.09 32.34 26.39 35.97 29.69 26.39 26. 70 32.87 29.12 29.70 51.51 56. 75 47.36 66.19 58.80 59.23 43.22 42.48 41. 12 30.04 44.44 37.77 30.04 30.28 37. 73 32.60 35.26 T1 —,.t-4 M c3 ~Z 0 C) 1 -M 38.90 43.13 34.47 42.37 30.99 34.31 39.27 43.05 30.96 34.49 New Hampshire...... New York........... Kansas.............. Michigan............ New Jersey.......... Pennsylvania....... Ohio................ 0.55 0.56 0.59 0.61 0.72 0.82 0.99 30.95 33.82 42.47 40.92 46.57 51.87 30.64 31.54 43.06 31.86 35.60 45.19 37.49 40.26 54.47 34.70 36.97 47.25 33.09 37.34 48.64 54.37 59. 35 55.16 60.43 65. 93 57.65 59.72 Alabama........... South Carolina....... Kentucky........... Tennessee.......... North Carolina....... Louisiana......... New Mexico......... 6.42 6.63 7.33 7.40 8.21 11.39 11.88 18.16 19.50 12.46 20.75 23.39 25.23 21.01 22.33 17.80 17.51 18. 40 21.55 10.02 24.86 26.93 30.58 24.87 29.39 30.44 34.98 29.49 35.14 25.71 30.59 30.94 33.86 31.05 53.01 *In cases of apparent tie, the percentages were carried out to three or four places. VFourteenth Census, vol. II, chap. XII, Table XII; Ayres, An Index Number for State School Systems, pp. 31, 33, 35, and 37. I,. t.4 z r) C) 0 0 t.4 C12 (/ 1-3 M3 V, 72 ILLITERACY IN THE UNITED STATES In conclusion, it may be stated that the general statement of the relation between illiteracy and school systems has been taken out of the realm of general statements and' by statistical methods been reduced to the significant coefficient of correlation, r= -.7. In other words, high illiteracy rates are found in those states whose school systems have been below the average in the past and which continue to rank comparatively low at present. CHAPTER VIII ADJUSTED ILLITERACY RATES The statistical tool, known as the method of standard population, is applicable to practically all sociological data which may be studied on the basis of age classifications. Since it makes it possible to hold the age distribution constant (i.e., the percentage of a given population that falls within a definite age period), the method of standard population permits a more accurate presentation of data than is usually given. The present chapter is devoted to presenting the applicability of the method as it applies to illiteracy data and to pointing out the changes in the alignment of states that result. Native whites of native parentage are studied in order that the race and "nationality" factor may be held constant. Table XX presents the percentages of illiteracy for this group according to the actual census data. The percentages were carried to the hundredth place for the sake of accuracy in dealing with the extremely low rates of illiteracy in certain states. Tht percentage age distribution of the entire native white population of native parentage, ten years of age and over, for the United States as a whole was taken as the standard. The total populations for the various states were then distributed on this basis. The numbers falling within the specific age periods for each state were multiplied by the illiteracy rates according to the actual distribution and the percentage of illiteracy for the states was then computed.l 1Delaware offers a good example of the method. Her population in 1920 contained a much higher percentage of old people than did the population for the 74 ILLITERACY IN THE UNITED STATES In some states it is evident that the population distribution is practically average as the illiteracy rates show slight change after the population distribution is adjusted. In Connecticut and Nebraska the rates remained the same to the hundredth place. In at least half of the states the adjusted distribution makes a relatively significant change in the percentage of illiteracy. Some of these changes increase the rates, others decrease the rates. A change of one-tenth of I per cent may well be regarded as significant where the rates of illiteracy are extremely low as such a change may actually make a 20 or 25 per cent difference in the given rate. Table XXI ranks the states according to the census figures of 1920 for illiteracy among native whites of native parUnited States as a whole which tended to raise her illiteracy rate. All figures are for native whites of native parentage. Percentage Population Adjusted Distribu- Distribu- Distribu Pr Cent Adjusted Age Period tionofU.S. tion for tion for Illiterate Number of Population Delaware Delaware 1920 Illiterates 1920 1920 (1) (2) (3) (4) (5) (6) 10-14 Years..... 14.68 13,094 16,373 0.2 33 15-24 Years..... 24.50 24,581 27,326 0.6 164 25-34 Years..... 20.14 21,952 22,463 1.0 225 35-44 Years..... 15.72 18,228 17,533 1.8 316 45-54 Years..... 11.34 14,450 12,648 3.2 405 55-64 Years..... 7.39 10,566 8,242 4.6 379 65 Years and Over 6.23 8,664 6,949 7.0 486 111,535 111,535 2,008 2,008 is to be compared with the actual number of illiterates in Delaware in 1920-2,290. The corresponding percentages of illiteracy are 2.05 and 1.80 (adjusted). ADJUSTED ILLITERACY RATES 75 I TABLE XX PERCENTAGE OF ILLITERACY AMONG NATIVE WHITES OF NATIVE PARENTAGE ACCORDING TO CENSUS DATA AND ACCORDING TO THE RESULTS OF THE METHOD OF STANDARD POPULATION, FOR 19202 Percent- Percent- Percent- Percentage of age of age of age of State Illiteracy Illiteracy State Illiteracy Illiteracy (Actual (Adjust- (Actual (AdjustData) ed Data) Data) ed Data). Daa_ Maine........... New Hampshire... Vermont......... Massachusetts.... Rhode Island..... Connecticut...... New York........ New Jersey....... Pennsylvania..... Ohio............. Indiana.......... Illinois........... Michigan......... Wisconsin........ Minnesota........ Iowa............. Missouri.......... North Dakota..... South Dakota..... Nebraska......... Kansas........... Delaware......... Maryland........ Virginia.......... 1.27 0.55 1.05 0.30 0.50 0.41 0.56 0.72 0.82 0.99 1.36 1.06 0.61 0.47 0.36 0.52 2.20 0.26 0.31 0.40 0.59 2.05 1.99 6.05 1.25 0.51 0.95 n in 0.46 0.41 0.54 0.56 0.79 0.94 1.24 1.08 0.61 0.56 0.40 0.41 2.18 0.30 0.34 0.40 0.58 1.80 1.94 6.09 West Virginia.... North Carolina.. South Carolina... Georgia......... Florida.......... Kentucky....... Tennessee....... Alabama........ Mississippi...... Arkansas........ Louisiana....... Oklahoma....... Texas........... Montana........ Idaho........... Wyoming........ Colorado........ New Mexico..... Arizona......... Utah............ Nevada......... Washington...... Oregon.......... California....... 4.83 8.21 6.63 5.48 3. 13 7.33 7.40 6.42 3.59 4.57 11.39 2.38 2.17 0.34 0.31 0.35 1.66 11.88 1.28 0.34 0.42 0.29 0.39 0.39 5.02 8.57 7.00 5.76 3.12 7.47 7.53 6.83 3.73 4.81 12.12 2.58 2.29 0.38 0.35 0.40 1.67 12.49 1.35 0.44 0.44 0.30 0.40 0.37 2Fourteenth Census, vol. II, chap. XII, computed from data given in Table XII. 76 ILLITERACY IN THE UNITED STATES TABLE XXI RANK OF STATES ACCORDING TO PERCENTAGE OF ILLITERACY AMONG NATIVE WHITES OF NATIVE PARENTAGE, ACCORDING TO CENSUS DATA, AND ACCORDING TO THE RESULTS OF THE METHOD OF STANDARD POPULATION. FOR 1920 Rank of Rank of Rank of States States Rank of States States According to Census Data* Accord- According to Census Data* According to ing to Adjusted Adjusted State Rank Data State Rank Data North Dakota...... Washington......... Massachusetts...... Idaho.............. South Dakota....... Utah............... Montana........... Wyoming........... Minnesota.......... California.......... Oregon............. Nebraska........... Connecticut......... Nevada............ Wisconsin.......... Rhode Island....... Iowa............... New Hampshire.... *New York.......... Kansas............. Michigan........... New Jersey......... Pennsylvania....... Ohio............... 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 1 3 2 5 4 14 7 8 11 6 9 10 12 15 19 16 13 17 18 20 21 22 23 24 Vermont.......... Illinois............ Maine............ Arizona........... Indiana........... Colorado.......... Maryland......... Delaware.......... Texas............. Missouri......... Oklahoma......... Florida............ M ississippi......... Arkansas.......... West Virginia...... Georgia........... Virginia........... Alabama.......... South Carolina.... Kentucky........ Tennessee......... North Carolina..... Louisiana.......... New Mexico....... 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 25 26 28 29 27 30 32 31 34 33 35 36 37 38 39 40 41 42 43 44 45 46 47 48 *In cases of apparent tie, the percentages on which the rankings are based were carried out to three or four places. ADJUSTED ILLITERACY RATES 77 entage. North Dakota with a rate of only 0.26 ranks first while New Mexico with a rate of II.88 is forty-eighth.3 When the ranking according to the adjusted data is given, various changes are noted. North Dakota retains first place while Washington drops from second to third place and Massachusetts rises to second place. Utah drops from sixth to fourteenth place; Oregon rises from eleventh to ninth, Nebraska from twelfth to tenth; Wisconsin changes from fifteenth to nineteenth; Iowa rises from seventeenth in rank to thirteenth. Most of the succeeding states retain their former ranking, even though there are significant increases and decreases in the actual percentages of illiteracy. It is believed that the computed percentages are a more accurate index of the actual illiteracy rates for the native whites of native parentage of the various states than any which have hitherto been presented. 3The data according to which the ranking was made are presented in the preceding table. CHAPTER IX CONCLUSIONS TO PART I The object of Part I has been an investigation of the general problem of illiteracy as it exists in the United States at the present time, together with the trend of illiteracy for the past fifty years. As a result of the investigation of the data certain conclusions present themselves. The preceding chapters contain analyses of the general results summarized below: I. The trend of illiteracy in the United States has been steadily downward, decreasing from an illiteracy rate of 20.0 per cent in 1870 to a rate of 6.0 per cent in I920. 2. The trend of illiteracy has been steadily downward in the nine geographical divisions of the United States from I870 to I920. The ranking of the divisions in I920 was as follows: Division Percentage of Illiterates in I920 West North Central. —. --- —------- 2.0 Pacific........ ------------ 2.7 East North Central ------------- 2.9 New England-..... --- —----— 4.. — 4.87 Middle Atlantic.... --- —---- 4.90 Mountain --------------- ----------- 5.2 West South Central. --- —------— Io.o South Atlantic. --- —-5 --- —----—. 5 East South Central -------------- 12.7 3. The trends of illiteracy in the various states from 1870 to I920 are to be largely explained in terms of age distri CONCLUSIONS TO PART I 79 bution, sex distribution, race and "nationality" groupings, percentage of urbanization, and school facilities. 4. Although the male illiteracy rate was lower than the rate for females from I870 to I9IO, in I920 the female illiteracy rate for the United States was 5.9 per cent as compared with a male illiteracy rate of 6.0 per cent. 5. Thirty-four of the forty-eight states had a higher percentage of male illiteracy than of female illiteracy in 1920. 6. The percentage of illiteracy is greater for females than for males in urban environments, while the reverse is true in rural environments. 7. The apparent difference in illiteracy rates by sex largely resolves itself into the sex distribution according to urban and rural environments, age distributions, race and nationality distributions, and lastly the institutionalized educational facilities for the sexes both in the United States and in Europe. When these factors are adequately taken into consideration, sex differences in regard to illiteracy practically disappear. 8. As would be expected, there is a diminishing percentage of illiteracy as one proceeds from the older age groups to the younger. 9. If no further efforts were made to reduce the illiteracy rate and provided conditions remained the same, it may be assumed that the illiteracy rate for native whites of native parentage would approach I per cent as those persons in the older age periods die. Io. The slightly smaller percentage of illiteracy among females in I920 is largely a function of the removal of larger percentages of illiterate older females than of illiterate older males from the population. 11. The urban illiteracy rate for the United States in I920 was 4.4. per cent as compared with a rural illiteracy rate of 7.7 per cent. 80 ILLITERACY IN THE UNITED STATES 12. In those states in which the urban illiteracy rate was higher than the rural illiteracy rate, there was a large proportion of illiterate foreign-born whites in the urban population as a dominant factor. 13. When the four-part division of (a) rural districts, (b) cities of from 2,500 to 25,000 inhabitants, (c) cities of from 25,000 to ioo100,000ooo inhabitants, and (d) cities of 100,oo000 or more inhabitants is utilized and when the factors of race and nationality are eliminated, the general trend of a lower illiteracy rate as one goes from the rural to the quantitatively most urbanized communities is apparent. 14. Within the limits of the available data, it is evident that the difference between urban and rural illiteracy rates is diminishing. 15. The illiteracy rate for the foreign-born whites was higher in 1920 than it was in i880, the year when information on this group was first available. i6. In 1920 the Indians had the highest rate of illiteracy of any recorded race or "nationality" group in the United States (34.9 per cent) while the Negro group ranked second with a rate of 22.9 per cent. 17. The illiteracy rate of 1920 for native whites of foreign or mixed parentage was lower than the illiteracy rate of native whites of native parentage. (o.8 per cent as compared with 2.5 per cent). I8. In the Middle Atlantic, the Mountain, and the Pacific Divisions, the Negro rate of illiteracy in 1920 was lower than the rate for native whites of native parentage in the South Atlantic and East South Central Divisions. 19. The foreign-born white group is the only one of the race and "nationality" groups in which the difference in illiteracy rates between urban and rural areas is slight. CONCLUSIONS TO PART I 81 20. The coefficient of correlation for illiteracy and school systems is at least, r = -.7 for each of the years studied. 2I. The method of standard population has been utilized in securing illiteracy rates for native whites of native parentage in such a way that actual conditions are more adequately presented than the unrefined data make possible. 4 PART II ILLITERACY IN RELATION TO CERTAIN SOCIAL PHENOMENA CHAPTER X THE RELATION OF ILLITERACY TO BIRTH-RATE The chief importance of the study of illiteracy as a societal phenomenon is its relation to other societal phenomena. It is the object of this and the succeeding chapters to utilize statistical and analytic techniques in the analysis of the relation of illiteracy to certain selected social factors. The methodology employed utilizes simple, partial, and multiple correlation in the endeavor to measure mathematically these several relationships. Regression equations are also presented as a basis of prediction. The use of statistical technique, properly guarded, places in the hands of the sociologist an instrument which enables him to approximate the manipulative methods of the more exact sciences. The first relationship to be considered will be that existing between illiteracy and the birth-rate. Births are the basis and the necessary prerequisite of a continuing population. Any factor which, when varied, affects variances in the birth-rate is of importance. Illiteracy as a societal phenomenon-as has been pointed out above-achieves importance as it affects other phenomena. The object of the present chapter is the measurement and the analysis of the relation of illiteracy to the tremendously important factor of birth-rate. The data utilized and the procedure involved follow. The rate of illiteracy per i,ooo women, of all classes, fifteen to forty-four years of age, and the number of legitimate births per i,ooo married women within the same age limits 7 86 ILLITERACY IN THE UNITED STATES were first computed. The data on births were utilized for twenty-one of the twenty-three states included in the birth registration area of the United States in I920. Utah was omitted because the high birth-rate was so clearly a departure from the normal due to the influence of Mormon culture. Maine was necessarily omitted due to the fact that births were not recorded in accordance with the age distribution of the mothers, and it was necessary to hold the important factor of age constant. In so far as possible only legitimate births were included though cases of unknown legitimacy were tabulated by the census as legitimate, while in California, Massachusetts, New Hampshire, and Vermont the birth certificates did not call for this information though it was sometimes given. The error, however, is slight, due to the facts that illegitimacy forms a small percentage of births, and that part of these illegitimate births are returned as such. The validity of the results obtained from computations based upon twenty-one states might perhaps be questioned. Each item, however, is not a single datum in itself but is an average obtained from hundreds or even thousands of individual cases. Thus the statistical universe is much larger than a cursory survey of the data would reveal. However, these twenty-one states may not be regarded as wholly representative of the United States as the registration area does not include a proportionately equal number of states from the various sections. The character of the results obtained would seem to justify the utilization of such data for I920 as are available. In general, these data appear reliable. In an analytic study of the relation of illiteracy to birthrate, certain important factors should also be included. That a definite relation exists between birth-rate and percentage of urbanization and between birth-rate and in *. -f. ILLITERACY AND BIRTH-RATE 87 come has long been known. Hence these two factors were included in the more refined computations and the character of the results, it is believed, justifies their inclusion. The data utilized for the four variables are presented in Table XXII. The notation employed in the correlations is as follows: Xa is the illiteracy rate for women of all classes, 15-44 years of age. X, is the number of legitimate births per i,ooo married women of all classes, I5-44 years of age. Xd is the per cent of urbanization for the total population. Xe is the per capita current income, averaged for i919, i920, and i92I. Current income rather than total income is utilized, not only because the current income figures are more reliable' but also because it would appear to be more satisfactory for the purposes of this study. Turning to the data, the linearity of the relation of illiteracy to birth-rate was first determined. The coefficient of correlation' secured for illiteracy and birth-rate when data for all classes were utilized was, rac= +.6i In order to refine the data by holding race and "nationality" constant, the data for native-born white women3 were next computed. The relationship again being linear, the coefficient of correlation was found to be r=+.75 This higher coefficient emphasizes the significant relationship which exists between illiteracy and high birth'Maurice Leven, Income in the Various States, chap. I. 2The errors for all coefficients of correlation are not included in the text to avoid possible confusion but are included in Appendix A, where the method of computation is also presented. 3See Appendix B. 88 ILLITERACY IN THE UNITED STATES TABLE XXII DATA UTILIZED IN THE CORRELATION OF ILLITERACY AND BIRTH-RATE, ALL CLASSES4 Illiteracy timate Rate Per BirthsPe Percentage Per 1,000 1,000 of Urbani- Capita State Females Married zation Current 15-44 Females Income -4 15-44 Xa X Xd Xs California.......... 30.9 132. 8 68.0 $ 909 Connecticut......... 72. 3 176. 2 67. 8 724 Indiana............ 12.6 149.9 50. 6 540 Kansas............. 10. 8 154.6 34. 9 546 Kentucky.......... 50.4 179.1 26.2 372 Maryland.......... 38.6 167.3 60.0 670 Massachusetts...... 51. 2 177. 7 94. 8 835 Michigan.......... 26.3 168.9 61. 1 659 Minnesota.......... 11.4 181.1 44.1 530 Nebraska........... 9.8 166.1 31.3 516 New Hampshire..... 42.9 174.6 63.1 612 New York.......... 53.3 157.2 82.7 943 North Carolina...... 96.6 224. 8 19.2 329 Ohio.............. 23.3 143.6 63.8 650 Oregon............. 9.4 126.3 49.9 685 Pennsylvania....... 45.5 178.7 64. 3 687 South Carolina...... 166.3 185.8 17. 5 312 Vermont........... 17. 5 161.3 31.2 544 Virginia............ 74.0 196. 7 29. 2 389 Washington......... 13.2 131.8 55.2 710 Wisconsin.......... 15.5 167.8 47.3 562 4Sources: Xa-Fourteenth Census, vol. II, chap. XII, computed from Table XII; Xc-Computed from Birth Statistics for 1920, Table V, and Fourteenth Census, vol. II, chap. IV., Table XI; X,-Fourteenth Census, vol. II, chap. I, Table XX; X,-Maurice Leven; Income in the Various States, p. 267. ILLITERACY AND BIRTH-RATE 89 rates, that is, that illiterate women tend to bear the larger number of children. At this point, further analysis requires the holding constant of the factors of urbanization and per capita current income in order to study their effects separately and in combination. The method of partial correlation was employed, data for all classes being utilized. The correlations of the zero order, the first order, and the second order are included in Table XXIII because of the light they throw upon the discussion, although it is obviously impossible to analyze each one separately. The correlations were calculated to the fourth place for algebraic reasons although they are only presented to two places. The accuracy of the correlations has been carefully checked and rac was calculated according to two methods as a further check. Tables by Truman Kelley5 and John Rice Miner6 were used to facilitate computation in this as in succeeding problems. TABLE XXIII GROSS AND NET CORRELATIONS rac= +.61 rd.e= +.20 rac.de= +.49 rad= -.27 rae.c=-.04 rad.ce= -.06 rae= -.38 rae.d= -.34 rae.cd= +.04 rcd=-.38 rd.a= -.28 rcd.ae= +.44 rce= -.59 rcd.e= +.47 Tce.ad=-.55 rd = +.91 rce. a= -.48 rde.ac =+.92 rac.d=+.57 re.d=-.63 rac.e= +.52 rde.a= +.91 rad.c -.06 tde.c= +.92 6Tables: To facilitate the calculation of Partial Coefficients of Correlation and Regression Equations. Tables of V 1-r2 and 1- r2for Use in Partial Correlation and in Trigonometry. Notation employed: a= Illiteracy rate for women 15-44 years of age; c=the number of legitimate births per 1,000 married women of all classes; d = percentage of urbanization for total population; e=current income per capita. 90 ILLITERACY IN THE UNITED STATES Whereas rac=+.6I, rac.d=+.57. As has been pointed out, the gross correlation of illiteracy and birth-rate appears significant. The correlation of these two factors, holding percentage of urbanization constant, while slightly lower, is still significant. This bears out the hypothesis that urbanization is a factor affecting the birth-rate and should be held constant in order to study the relation of illiteracy to birth-rate more precisely. Furthermore, as was pointed out in Chapter VI, rates of illiteracy tend to be higher in rural than in urban areas. As a check on this analysis the correlation between birth-rate and urbanization is found to be rcd = -.38 while the correlation between illiteracy and urbanization is rad= -.27. Turning to the factor of per capita current income, the gross correlation of +.6I becomes a correlation of +.52 (rac.e= -.52). The importance of per capita current income is apparent. The simple correlations involved are rae= -.38 and rce=-.59. Carrying the analysis a step further, the relation between illiteracy and birth-rate is obtained when both urbanization and per capita current income are held constant. The correlation when these two factors, both of which are negatively correlated with illiteracy and birthrate, are assigned is further lowered so that rac.de= +.49. This correlation while lower than rac is actually more significant for the present purpose because two important factors are held constant. That is, when the effect of urbanization and the effect of current income are controlled, the relation between illiteracy and birth-rate may be stated in terms of the coefficient of correlation as+.49. Considering the fact that this is a net correlation, it assumes more significance, for the purposes of the present investigation, than the gross correlation of +.6I. Thus illiterate women tend to have higher birth-rates than literate women. ILLITERACY AND BIRTH-RATE 91 In general, investigators of the factors affecting birthrate have neglected the influence of illiteracy. Supporting the present conclusion, however, is the following statement in regard to English conditions. "It is worthy of note that the fall of the birth-rate in this country practically dates from the passing of the Education Act in 1 870. "7 Hornell Hart in his study of "Differential Fecundity in Iowa" states that the poorly educated are more fecund than the better educated.8 Del Vecchio has shown that for Italy the birth-rate is highest in the districts which contain the largest number of persons who have not learned to read.9 In general, where educational status has been used at all, the approach to the problem has been an effort to determine the relation of education to lowered birthrate. The present approach has been to determine instead the relation of illiteracy to high birth-rates. Thus this study is complementary to studies of the other type. Obviously there is no value in attempting to predict the illiteracy rate on the basis of the birth-rate and the additional factors studied. It is important to predict the birthrate, given a certain illiteracy rate. Interest chiefly centers around the prediction of the birth-rate when the illiteracy rate falls to a theoretical zero. The gross regression equation is Xc =.373Xa+151.33 with a corresponding error of estimate, Sca= I7.67. According to this equation, when an illiteracy rate of zero is substituted for Xa, the birth-rate would become 151.3 = 17.67 per 1,000 married mothers, 15-44. The net regression equation, however, offers a more accurate basis forprediction. Xc= 201.12+.237Xa+.823Xd -.142 Xe. When Xd and Xe are held constant at the 7J. Anderson, "The Falling Birth-Rate," Nature, XCI, p. 84. 8University of Iowa Studies in Child Welfare, vol. II, no. 2. p. 35. 9Quoted in L. Brentano, "The Doctrine of Malthus and the Increase of Population During the Last Decades," Economic Journal, XXIX (1910), 382. 92 ILLITERACY IN THE UNITED STATES mean and zero is again substituted for Xa, the value of Xc becomes 157. This is a more precise estimate of the birthrate as is indicated by the corresponding value of S. (Sc.ade=3.oo005). The mean of the given birth-rates is I66.8 while the mean for the illiteracy rate is 4I.5. The value of these predictions as well as the reliability of the partial correlations can easily be over-estimated. The reason is that the data are available for only a limited number of states. In spite of this limitation, however, the results appear sufficiently important to be considered as indicating definite trends. Hence it seems warrantable to state general conclusions on the basis of the results which have been secured from the application of certain statistical techniques to the data originally presented. The decreasing birth-rate is a phenomenon that is of tremendous importance. Various factors have been found partially to account for this. The economic factor has been given due weight in previous studies and the results of this chapter support such conclusions. Some students have noted the effect of urbanization upon the birth-rate, while few have noted the effect of illiteracy'~ or of low educational status. The present chapter was designed to test this hypothesis that illiteracy significantly affects the birth-rate but the result must only be considered a tentative one, subject to further investigation. To sum up briefly, one may say, then, that not only were low birth-rates found to be related to higher economic status, and to a high percentage of urbanization, but also to low illiteracy rates. Moreover, since urbanization and economic status were thought to be factors that disturbed the truer relationship of birth-rate 10Whenever the general term of illiteracy is used, it refers, of course, not to the actual number of illiterates but to the specific illiteracy rate as defined in the list of variables presented at the beginning of each chapter. ILLITERACY AND BIRTH-RATE 93 and illiteracy, these two affecting factors were mathematically controlled. When this was done, a significant relationship between illiteracy and birth-rates was still found to exist. It would appear, then, that illiterate women tend to bear more children. To what degree this is due to lack of knowledge of birth control in its various forms, to what degree it is due to lack of other sources of interest, is a separate study in itself and need not be entered into here. If this relationship be true then, with the gradual decrease of illiteracy, and conversely, with the gradual increase in the educational status of women, it would appear that we have here an important factor in explaining the decreasing birth-rate in the United States. While mathematical prediction of the birth-rate, given a certain illiteracy rate, is statistically possible, nevertheless the results must be viewed with caution due to the paucity of the data. Yet the trend is definitely apparent. As the trend of illiteracy in the United States is downward, one may venture the prediction that, if other factors remain the same, the birthrates of American families will also continue to tend downward. Increasing urbanization and the apparently increasing per capita current income will also work in the same direction, as the chapter indicates. Lastly, in connection with the results secured, it would appear that a complete study of the factors affecting birth-rates should certainly include illiteracy or some socio-educational equivalent. CHAPTER XI THE RELATION OF ILLITERACY TO INFANT MORTALITY Death, like birth, is a process which though biological in character, is largely conditioned by societal phenomena. The first year of life, the period covered by the term " infant mortality," is admittedly hazardous. If the illiteracy rate affects the infant mortality rate, its importance is perhaps obvious. The present chapter undertakes to measure as well as to analyze, the relation of illiteracy to infant mortality in the United States. Infant mortality refers to the ratio of deaths of infants under one year of age per i,ooo births, exclusive of stillbirths. Thirty-three states were utilized in the study. The infant mortality rate for twenty-three states was obtainable for I920, while six additional states were obtainable for I92i or i922. Three states were included in the registration area for the first time in 1924, and one, West Virginia, was admitted in i925 (see Table XXIV). These latter four states show returns which are normal. Birth and infant death reporting has been somewhat inaccurate in the past, but the states represented in the birth registration area are believed to have sufficiently accurate data, particularly for the period included in the scope of this study. The errors in reporting the age of a child do not affect the data, so long as they do not run past the one year age limit. The variables under consideration, therefore, are illiteracy rates for all classes, IO years of age and over,' and 'Census of 1920, vol. II, chap. XII, Table XII. ILLITERACY AND INFANT MORTALITY 95 TABLE XXIV DATA UTILIZED IN THE CORRELATION OF ILLITERACY AND INFANT MORTALITY, ALL CLASSES2 Per Cent I- Deaths of In- Per literate, 10 fants under 1 ercentage Capita Years of Age Year of Age of Urbani- Current State and Over Per 1,000 zation Income and Over Income Births Xf X, Xd Xe California.......... 3. 3 74 68.0 909 Connecticut....... 6. 2 92 67. 8 724 Delaware.......... 5. 9 98 54. 2 656 Florida............. 9. 6 82 36. 7 402 Illinois............. 3.4 76 67.9 764 Indiana........... 2.2 82 50.6 540 Iowa............... 1.1 55 36.4 524 Kansas............. 1.6 73 34.9 546 Kentucky.......... 8.4 73 26. 2 372 Maine............. 3.3 102 39.0 569 Maryland.......... 5.6 104 60.0 670 Massachusetts..... 4. 7 91 94.8 835 Michigan........... 3.0 92 61.1 659 Minnesota.......... 1.8 66 44.1 530 Montana........... 2.3 70 31.3 573 Nebraska........... 1.4 64 31.3 516 New Hampshire..... 4.4 88 63. 1 612 New Jersey........ 5. 1 74 78. 4 745 New York.......... 5.1 86 82.7 943 North Carolina...... 13.1 85 19.2 329 North Dakota....... 2.1 67 13. 6 426 Ohio............... 2.8 83 63.8 650 Oregon............. 1.5 62 49.9 685 Pennsylvania....... 4.6 97 64.3 687 Rhode Island....... 6.5 93 97. 5 783 96 ILLITERACY IN THE UNITED STATES TABLE XXIV-Continued Per Cent I- Deaths of In-Per fants under 1 Percentage literate, 10 fants under 1 ercentage Capita Years of Age Year of Age of Urbani- Current State and Over Per 1,000 zation Inme and Over BIncome Births Xf XgXd Xe Utah............... 1.9 71 48.0 $ 514 Vermont........... 3.0 96 31.2 544 Virginia............ 11.2 84 29.2 389 Washington......... 1. 17 66 55.2 710 West Virginia....... 6.4 80 25.2 474 Wisconsin.......... 2.4 77 47.3 562 Wyoming.......... 2.1 79 29.5 826 South Carolina...... 18.1 116 17.5 312 Mean.............. 4.7 81.8 49. 1 $ 605 deaths of infants under one year of age per I,ooo births for all classes. The data are linear, and the coefficient of correlation is rig= +.57. This gross correlation coefficient appears significant. The utilization of the method of partial correlation is the next step in the analysis of the data. As in Chapter X, the percentage of urbanization and the per capita current income would appear to be two factors of sufficient importance to affect the relationship between illiteracy and infant mortality. 2Sources: Xf-Fourteenth Census, vol. II, chap. XII, Table XII; Xg-Birth, Stillbirth, and Infant Mortality Statistics for the Birth Registration Area of the United States, 1924, pp. 31-32, and for 1925, p. 6; X, —Fourteenth Census, vol. II, chap. I, Table XX; Xe-Leven, Income in the Various States, p. 267. ILLITERACY AND INFANT MORTALITY 97 The notation utilized in the partial and multiple correlations is as follows: f=illiteracy for all classes, IO years of age and over. g = deaths of infants under I year of age per I,OOO births (exclusive of stillbirths.) d=percentage of urbanization. e=per capita current income. Table XXV gives the correlations of the zero, first, and second orders. TABLE XXV GROSS AND NET CORRELATIONS r/v-= +.57 rgd.f = +.37 rf d =-.22 rd.e= +-.28 r/e=-.45 r,,.=+-.37 rod= +.18 rge.d= -.21 r,e= +.02 rde f= -.82 rde= +.81 rde.g= +.82 r/7.d= +.63 rfo.de = +.61 rj.e= +.64 r/d. e= +.14 r rf., -.39 rfe.gd= -.45 r/d.e=+.28 rg<d./e=+.13 rfe.= -.56 roe.fd=+.12 rfe. = -.47 rde.og +.79 The gross correlation between illiteracy and infant mortality as has been stated, is +.57. Holding the factor of percentage of urbanization constant, the coefficient of correlation becomes +.63 (rf.d= +.63). Consideration of the correlations of the zero order shows that while the correlation between illiteracy and urbanization, rfd= -.22, is negative, the correlation between infant mortality and urbanization, rd= +.I8, is positive. Thus when the factor of urbanization which is differently cor3Table XXIV contains the data on which the computations were based. 98 ILLITERACY IN THE UNITED STATES related with the two primary factors is held constant, the net relationship is found to be higher than the gross correlation would indicate. When the factor of per capita current income is controlled, rfg.e= +.64. The gross correlation of illiteracy and current income is negative (rfe = -.45) while the gross correlation of infant mortality and current income is practically zero (r,= +.02). Therefore, when the influence of per capita current income on these two variables is controlled, the correlation between illiteracy and infant mortality rises significantly. When both percentage of urbanization and per capita current income are held constant, the correlation is significantly high (rf,.de= +.6I). Thus it may be stated that the gross relationship between illiteracy and infant mortality is a significant one; but when the two important factors of percentage of urbanization and per capita current income are controlled, the refined relationship is found to be even more significant. The most important studies, for purposes of comparison, are those made under the direction of the Children's Bureau. These studies, while important individually for their corroboration of the relation of illiteracy and infant mortality, increase in significance as they support each other. In a field study made in Baltimore, Maryland, the infant mortality rate for native white literate mothers was 94.1 per I,ooo births as compared with I89 per I,ooo births for native white illiterate mothers.4 In a similar study conducted in Johnstown, Pennsylvania, the infant mortality for literate foreign mothers was found to be I48.0 as compared with 214.0 for illiterate foreign mothers.5 In Mont4Anna Rochester, Infant Mortality, Results of a Field Study in Baltimore, Md., Based on Births in One Year, p. 332. 5Emma Duke, Infant Mortality, Results of a Field Study in Johnstown, Pa., Based on Births in One Calendar Year. ILLITERACY AND INFANT MORTALITY 99 clair, New Jersey, it was found that the infant mortality rate where mothers were literate, was 76.9 while the rate where mothers were illiterate, was o08.4.6 Similar studies were conducted in Pittsburgh, Pennsylvania, in Waterbury, Connecticut, in Akron, Ohio, in Brockton, Massachusetts, in Manchester, New Hampshire, and in Saginaw, Michigan, where approximately. similar results were secured. On the basis of the results previously secured, prediction of the rate of infant mortality, given a specified illiteracy rate, is important. The gross regression equation is X,=2.052Xf+72.07 while Sgf=II.I2. More precise results may be secured by utilization of the net regression equation, X= 55.04+2.488Xf+.II3Xd+. oI56Xe. The standard error of estimate becomes Sg.fde= 10.23. Therefore, when urbanization and per capita current income are held constant at the mean and the illiteracy rate is theoretically reduced to zero, the infant mortality rate becomes 70.1 per I,ooo live births, for the total registration area. This must be regarded as only a tentative figure inasmuch as all the states were not included in the registration area when the data were computed. Moreover, the states included are, in general, those which do not have the highest rates of infant mortality. The general relationship of high illiteracy rates with high infant mortality is evident. The first year of life is the most hazardous. The infant mortality rate covers this period. With the human organism so poorly equipped to survive, the importance of affecting factors is obvious. Illiteracy, particularly on the part of the mother, seems to be a factor affecting continuance of infant life. Apparently, at least as important as the economic factor is the factor of illiteracy. This latter 6Children's Bureau; U. S. Dept. of Labor. A Study of Infant Mortality in a Suburban Community, Montclair, N. J., p. 23. 100 ILLITERACY IN THE UNITED STATES is definitely contributory to the unsuitability of many mothers to adequately provide the necessary safeguards for their offspring. The relationship, of course, is not perfect; yet it would seem to be of tremendous importance. Ignorance in this matter can be successfully overcome by both private and public child welfare agencies. The present chapter deals with the thirty-three states included in the registration area in I925, data of non-registration states, even where available, being rejected on the score of doubtful validity. The results show a significant relationship between illiteracy and high infant mortality, a result to be expected if the foregoing relationship holds. The excellent studies of the Children's Bureau corroborate the findings of this chapter through their more intensive studies of smaller areas. Prediction of the future so far as the factors under consideration are concerned may be made, subject to a certain margin of variance. In so far as society is concerned, the outlook is hopeful, for as the trend of illiteracy continues downward, it would appear that the infant mortality rate should be reduced. Finally, it is evident, if the above holds true, that no analysis of the social factors affecting infant mortality can be complete unless illiteracy or some socio-educational equivalent is adequately taken into account. CHAPTER XII THE RELATION OF ILLITERACY TO EARLY AGE OF MARRIAGE The object of this chapter is the measurement and analysis of the relation of illiteracy to early age of marriage, The rate of illiteracy for native whites of native parentage. ten years of age and over, is utilized as representing the former factor.' Early age of marriage is obtained by securing the ratio of females I5-I9 ever married2 to the total female population I5-I9, native whites of native parentage being utilized.3 Females are taken rather than males because the data for males are not as reliable and due to the nature of the investigation only one sex is necessary. The age period, 15 to I9, is utilized because prior to 15 years there is such a small percentage of marriages. The statistical universe is the forty-eight states. Before proceeding with the study of the relation of illiteracy to early age of marriage for native whites of native parentage, the relationship of these factors, taking all classes into consideration,4 was measured. The Pearsonian coefficient of correlation was found to be r= +.64, indicating a significant relationship. The data being slightly nonlinear in nature, the index of correlation, or Rho, was computed on the basis of a second degree parabola.5 The slight'Fourteenth Census of the United States, vol. II, chap; XII, Table V. 2Those women who are married, widowed, or divorced are under the same classification, inasmuch as they have all entered the marriage state. 3Fourteenth Census of the United States, vol. II, chap. IV, Table XI (computed). 4See Appendix B for data utilized. 5The formula used was a z(Y)+b I2(XY)+c Z(X2Y) -Nc2 Rho2,(z — Nc 2:( Y) - Nc,2 102 ILLITERACY IN THE UNITED STATES ly better fit to the data was evidenced by the fact that the index of correlation was found to be Rho =.68 in comparison with the previously computed correlation coefficient, r= +.64. Following this generalized picture of the relationship of illiteracy to early age of marriage, the more controlled group of native whites of native parentage was studied. These data were tested and found to be linear. The correlation was found to be r=+.63, which may be considered a significant one. This, however, is affected by other factors, and calls for the utilization of partial correlation methods in order to present a truer picture of the subject. As in the preceding two chapters, percentage of urbanization and current income per capita were selected as the two most important measurable factors (in addition to illiteracy) affecting early age of marriage, and hence to be controlled. The percentage of urbanization6 for native whites of native parentage ten years of age and over was selected as the third variable, because it was believed that the age factor should be controlled in the variable to this extent. The fourth variable, current income per capita, remains the same as in preceding chapters because of the impossibility of further refining this variable. However, it is utilized because of the fact that this carefully worked out index of current income is, as presented, possibly the most important indicator of economic wealth which in turn affects the age of marriage. The notation utilized in the partial coefficients of correlation is as follows: b=illiteracy rate of native white of native parentage, Io years of age and over. 6Computed from the Fourteenth Census, vol. II, chap. I, Table XX, and chap. II, Table XIII. ILLITERACY AND EARLY MARRIAGE 103 TABLE XXVI DATA UTILIZED IN THE CORRELATION OF ILLITERACY AND EARLY AGE OF MARRIAGE, NATIVE WHITES OF NATIVE PARENTAGE7 Per Cent Percentage Percentage of Per Illiterate, of Females, Urbanization Capita State 10 Years of 15-19 Ever 10 Years of Current Age and Over Married Age and Over Income Xb Xh Xi X. Maine.............. 1.3 11.5 31.7 $ 569 New Hampshire..... 0.6 10.1 49.1 612 Vermont............ 1.1 10. 1 26. 2 544 Massachusetts...... 0.3 4.7 90.3 835 Rhode Island....... 0.5 5.5 93.9 783 Connecticut......... 0.4 6. 2 57.9 724 New York.......... 0.6 7.3 67. 1 943 New Jersey......... 0. 7 8.0 68. 4 745 Pennsylvania....... 0.8 9. 7 56.3 687 Ohio............... 1.0 11.6 55.0 650 Indiana............ 1.4 13.6 45.5 540 Illinois............. 1. 1 10. 8 52. 1 764 Michigan........... 0.6 12.7 55.2 659 Wisconsin........... 0.5 5.6 44.1 562 Minnesota.......... 0.4 6.1 47.0 530 Iowa............... 0.5 10. 5 38.3 524 Missouri.............. 2. 14. 1 38.9 542 North Dakota....... 0. 3 7. 6 20. 2 426 South Dakota....... 0.3 9.3 21.3 500 Nebraska............ 0.4 10. 7 3,2.6 516 Kansas............. 0.6 12.6 36.2 546 Delaware........... 2.0 13.7 45.9 656 Maryland.......... 2.0 12.4 54.6 670 Virginia............ 6.1 13.8 28.6 389 West Virginia....... 4.8 17. 7 25.0 474 North Carolina...... 8.2 16. 3 19. 6 329 South Carolina...... 6.6 17.3 22.1 312 Georgia............. 5.5 20.9 27.5 722 Florida............. 3.1 19.6 32.4 402 104 ILLITERACY IN THE UNITED STATES TABLE XXVI-Continued Per Cent Percentage Percentage of Per Illiterate, of Females, Urbanization Capita State 10 Years of 15-19 Ever 10 Years of Current Age and Over Married Age and Over Income Xb Xh Xi Xe Kentucky.......... 7.3 20. 6 22.2 $ 372 Tennessee.......... 7.4 19.5 23.6 336 Alabama........... 6.4 20.8 21.6 291 Mississippi.............3.6 17.3 16.6 263 Arkansas............ 4.6 22.4 17.3 306 Louisiana............ 11.4 16.1 36.0 395 Oklahoma.......... 2.4 21.6 28.1 466 Texas.............. 2.2 17.6 32.7 497 Montana........... 0. 3 11.6 31.4 573 Idaho.............. 0.3 13.5 28.9 549 Wyoming........... 0.4 15.4 28.7 826 Colorado........... 1. 7 14.2 48.6 670 New Mexico........ 11.9 18.0 18.8 445 Arizona............ 1.3 17.9 40.4 640 Utah............... 0. 3 10.0 46.9 514 Nevada............ 0.4 13.7 23.6 874 Washington........ 0.3 14. 6 54.2 710 Oregon............. 0.4 14. 1 47.3 685 California.......... 0.4 11.4 68.1 909 M ean.............. 2.4 13.3 40.0 $ 564 Sources: Xb-Fourteenth Census, vol. II, chap. XII, Table V; Xh-Ibid., chap. IV, computed from Table XI; Xi-Ibid., computed from chap. I, Table XX and chap. II, Table XIII; Xe-Leven, op. cit., p. 267. h=per cent of females, 15-19, ever married, native white of native parentage. 7Except data on per capita current income which is necessarily for all classes. ILLITERACY AND EARLY MARRIAGE 105 =per cent of urbanization, native white of native parentage, IO years of age and over. e = current income per capita. All the correlations of the zero order, the first order, and the second order are to be found in Table XXVII. TABLE XXVII GROSS AND NET CORRELATIONS rh = +.63 rhi.b = -.50 bi = -.49 rhi. = -.38 rbe= -.65 re., = -.33 rhi = -.65 the.i = -.23 rhe =-.60 rie.b = +.65 rie = +.75 rie.h = +.59 rbh.i = +.48 rbh.i. = +.43 rbh.e,= +.40 fbi.he = +.17 rbi.h = -.13 rbe.hi = -.44 rbi.e= +.01 rhi.be = -41 rbe.h = -.43 rhc.bi = -.0006 nbe.i= -.49 ri,.bh = +.60 The gross correlation of illiteracy and early age of marriage is +.63, which is indicative of an important relationship, to be further tested. When the factor of urbanization is held constant, the correlation is rbh.i=+.48, a significant correlation, with the influence of urbanization mathematically controlled. The correlation between illiteracy and urbanization is found to be rbi=-.49, showing the important negative relationship of illiteracy and urbanization. The correlation between urbanization and early age of marriage is rhi= -.65, showing the importance of the former factor which was controlled in obtaining the correlation, rbh.i= +.48. The gross correlation of +.63 was then refined by holding the current income per capita constant, the resulting 106 ILLITERACY IN THE UNITED STATES correlation being rbh.e= +.40, a not particularly high, yet statistically significant, correlation. The gross correlation of rb,=-.65 between illiteracy and current income per capita, and the gross correlation of rhe=-.60 between early age of marriage and current income per capita emphasize the important negative relationships which influence the simple correlation, rbh= +.63. Although both affecting factors are so clearly important, the relationship between illiteracy and early age of marriage, holding both percentage of urbanization and current income per capita constant, appears significant. rbhie= +.43..In their study of the families of Mid-Western college students, Baber and Ross found that for men of the "present generation," "While education does have a retarding effect on marriage, the postponement from this cause is not very great."8 For women, also of the "present generation," "As they climb the educational stairs the postponement of their wedding day becomes not only absolutely but also relatively later as though the stairsteps were of increasing height."9 When it is desired to predict the percentage of married females, 15-I9, from a given illiteracy rate, the gross and net regression equations are as follows: Xh =.988X+ 10.93 Xh = I6.40+.645Xb-. I I 56Xi —.ooooI 3X. The standard errors of estimate are Shb= 3.57 and Sh.bie= 3.08. When percentage of urbanization and current income per capita are held constant at their means and zero is substituted for Xb in the net regression equation, II.77 is 8Ray E. Baber, and Edward A. Ross; Changes in the Size of American Families in One Generation. p. 31. 'Ibid., p. 32. ILLITERACY AND EARLY MARRIAGE 107 secured as the percentage of married females, 15-19, native white of native parentage. Thus it is evident that there are a number of social factors which affect the age at which populations marry. While the sociologist who looks at the family from a long range standpoint has no fears but that some form of family life will continue to exist, it must be noted that there are various factors which tend to compete, in early years at least, with entrance into the family relationship. Urbanization is an example of this, offering as it does various alternatives to marriage. Higher economic status would also appear to be an important factor with its various intellectual inducements to at least postpone marriage. The results of the present chapter confirm both of these hypotheses, there being significant negative correlations between early age of marriage, on the one hand, and the factors of urbanization and economic status, on the other. But the main concern here is with the relationship of illiteracy, as an index of low educational status, to the phenomenon of early marriage. A significant correlation (+.63) is found, and when the highly important correlated factors of urbanization and current income are controlled, the mathematical relationship, while naturally lower, is of sufficient importance to be noted. When both factors are controlled, the resulting correlation is still important. One appears warranted in stating, then, that the illiterate woman, the woman of lowest educational status, tends to marry earlier than her literate sister. To what degree this is due to circumscribed environment brought about in part, or concomitant with, little education, (formal or otherwise) is interesting to speculate upon. Our concern is with the facts, and the correlation shows not only an important relationship, but also the use of the re 108 ILLITERACY IN THE UNITED STATES gression equation enables us to predict that the raising of the educational status of women will be one factor tending to postpone the age at which they marry, even though other factors, such as birth-control, for instance, may tend to operate in the opposite direction. Presumably, then, illiteracy or its equivalent is a factor to be definitely taken into account in any study of early age of marriage although its importance has apparently, in most cases, been overlooked by previous investigators in this field. CHAPTER XIII THE RELATION OF ILLITERACY TO SIZE OF FAMILY The object of the present chapter is the measurement and analysis of the possible relationship of illiteracy to size of family. On empirical grounds it would appear that such a relationship does exist. "Size of family," as such, is not computed by the Census. However, the average number of living children of mothers of I920 is available for twenty states. A preliminary investigation of this problem was made as follows. "Size of family" was approximated through obtaining the average number of living children of native white mothers of I920 for those twenty states of the registration area for which such information was obtainable.' This was correlated with the illiteracy rate per I,000 native white women, I5-54-2 The resulting Pearsonian coefficient of correlation was +.72.3 The correlation is apparently significant, but one may question the degree to which the factor "average number of living children of native white mothers of 1920" is sufficiently representative of the sought factor, "size of family." A second and perhaps more fruitful approach (in addition to possible corroboration of the above results) was 1Data for Maine and New Hampshire are not available for "children living." 2Computed from data obtained in U. S. Census, 1920, vol. II, chap. XII, Table XII. The data on which the computations were based is included in Appendix B. 3 r= +.70, when the illiteracy rate per 1,000 native white women, 15-44, was used. 110 ILLITERACY IN THE UNITED STATES determined upon. With the illiteracy rate per i,ooo native white women I 5-54, was correlated the number of children under I5 years of age per native white woman, ever married, I5-54.4 This approximation to " size of family "5 was utilized because it gives what is believed to be a more accurate picture of actual conditions. The broad age span for women ever married was taken so as to include not only mothers of the child-bearing period but also mothers of fairly advanced years who still have young children. If mothers 15 to 44 had been utilized together with all children under fifteen years of age, the number of children per married woman would have been larger than is actually the case. The gross coefficient of correlation is +.62. This is significant, and indicates a high degree of relationship, and in addition this is corroborative of the results of the incomplete sample utilized in the preceding correlation. One may infer, then, that there is an apparently important degree of relationship between illiteracy and the number of children under fifteen years of age, per I,ooo native white women, ever married, I5-54. The factors of current income and urbanization may partially account for this apparently high degree of relationship, however, and hence the method of partial correlation is again utilized to keep these factors, separately and together, constant. The notation employed in the following partial and multiple correlations is as follows: Xj=illiteracy per i,ooo native white women, I5-54. Xk=number of children under fifteen years of age per native white women, ever married, I5-54. 4U. S. Census, 1920, vol. II, chap. III, computed from Table I. 'Males were not included as obviously the number of ever married native white males would not correspond numerically to the same classification of females. ILLITERACY AND SIZE OF FAMILY 111 TABLE XXVIII DATA UTILIZED IN THE CORRELATION OF ILLITERACY AND SIZE OF FAMILY, NATIVE WHITES6 No. of ChilNo. Illiterate drenUnder 15 Percentage of Per Per 1,000 Na- Years Per Urbaniza- Capita tive White Ever Married tion of Native Current State Women Native White White Income Income 15-54 Woman Persons 15-54 Xi Xk Xi X, Alabama........... 50.5 2.14 21.1 $ 291 Arizona............ 24.8 1.35 39. 7 640 Arkansas........... 31.5 2.06 16.7 306 California.......... 3.6.88 68.4 909 Colorado........... 15.6 1.29 47.5 670 Connecticut......... 3.8.96 65.4 724 Delaware........... 9.9 1.26 52.6 656 Florida............. 22.0 1.63 34.5 402 Georgia............ 39.8 2.03 26.4 322 Idaho.............. 2.3 1.74 27.5 549 Illinois............. 4.6 1. 22 62. 7 764 Indiana............ 6.7 1.38 47.9 540 Iowa.............. 2.7 1.42 35.7 524 Kansas............. 3.6 1.52 33.5 546 Kentucky........... 52.1 1.91 23.6 372 Louisiana........... 106.2 1.92 38. 5 395 M aine............. 11.0.63 36.3 569 Maryland.......... 10.8 1.42 59.6 670 Massachusetts.... 3.3.98 93.9 835 Michigan........... 4.2 1.19 57.8 659 Minnesota.......... 3.3 1.34 42.5 530 Mississippi......... 24. 8 2. 13 16. 2 263 Missouri.......... 11.3 1.44 42.9 542 Montana........... 2.2 1.36 31.1 573 Nebraska........... 2.8 1.50 29.5 516 Nevada............ 3.1 1.06 21.8 874 New Hampshire..... 6.0 1.08 59.2 612 New Jersey......... 4.5 1.07 76. 3 745 112 ILLITERACY IN THE UNITED STATES TABLE XXVIII-Continued No. of ChilNo. Illiterate dren Under 15 Percentage of Per Per 1,000 Na- Years Per Urbaniza- Capita tive White Ever Married tion of Native Current State Women Native White White Income 15-54 Woman Persons 15-54 Xj Xk Xi X,.... New Mexico........ 145.7 1.93 18.8 $ 445 New York.......... 3.7.98 78.7 943 North Carolina...... 66. 4 2.30 18.6 329 North Dakota...... 3.6 1.54 14.1 426 Ohio............... 4.9 1.27 60. 3 650 Oklahoma.......... 15.5 1.89 26. 5 466 Oregon............. 2.3 1.18 48.4 685 Pennsylvania....... 5.2 1.38 61.6 687 Rhode Island....... 7.3.94 96.9 783 South Carolina...... 60.0 2.19 21.3 312 South Dakota....... 3.2 1.59 16.6 500 Tennessee............ 53.1 1.94 22.9 336 Texas.............. 25.9 1.75 31.5 497 Utah............... 2.4 1.91 47.0 514 Vermont........... 8.9 1.31 29.8 544 Virginia............ 41.6 1.96 28.1 389 Washington.......... 1.9 1.12 54.2 710 West Virginia....... 35. 7 2.04 24.8 474 Wisconsin.......... 4. 5 1.41 45. 7 562 Wyoming........... 2.2 1.42 28.8 826 Mean.............. 20.0 1.50 41.3 $ 564 Sources: Xi-Fourteenth Census of the United States, vol. II, chap. XII, computed from Table XII; Xk-Ibid., chap. III, computed from Table I; XI-Ibid., chap. I, Table XX; X,-Leven, op. cit., p. 267. 6With the exception of per capita current income which is necessarily for all classes. ILLITERACY AND SIZE OF FAMILY 113 XI= percentage of urbanization of native white persons. Xe=per capita current income. Table XXIX gives the correlation coefficients of the zero, first, and second orders. rjk= +.62, while rjk.l= +.57 and rik. e= +.35 -TABLE XXIX GROSS AND NET CORRELATIONS rik = +.62 rkl. j = -.68 ri = -.42 rkl.o = -.28 ri, = -.55 rke. -.78 rkl= -.-74 rk.I- -.75 rke = -.86 r., = +.69 rie = +.75 r,.k = +.23 rik./ = +.57 rjk.l'= +.36 rtk.e= +.35 ril.k.= +.11 ri.k = +.09 rie.kl = -.08 r ji.= -.002 rkl.,.= -.30 rijek = -.04 rk,. l = -.60 rie.l = -.40 rle.ik = +.34 The interpretation of these statistical results is that there is an important gross correlation between illiteracy and the variable utilized for size of family.7 When urbanization is held constant, the correlation is still important. When current income per capita is held constant the correlation, while a low one, is considered significant. The checks, rkl=-.74 and rke=-.86, show the necessity of holding the Xi and Xe factors constant since they show important negative correlations with size of family. In addition, rj= —.42 and re,= -.55, indicating important negative correlations between Xj and Xi and between Xj and Xe. These correlations also indicate the necessity of hold7Hereafter this variable will be called, for brevity's sake, "size of family." 114 ILLITERACY IN THE UNITED STATES ing percentage of urbanization and per capita current income constant. When either of these two significant factors is assigned, the net correlation between illiteracy and size of family is lowered. When both percentage of urbanization and per capita current income are held constant, rjk.le=+.36. This indicates a tendency for high rates of illiteracy and large families to be found together, for the contributing factors of urbanization and current income are mathematically controlled. Various studies have noted the fact that there is a relationship between size of family and educational status. S. J. Holmes found that "education has a potent effect in reducing family size."8 Better educated women tend to have fewer children than do those with less education, according to the findings of Mary R. Smith.9 Baber and Ross also found that on the average the more poorly trained groups have the larger families.~0 In addition, many of the studies of economic status and size of family show small families for those occupations which require the greater educational preparation. One may utilize regression equations for prediction of size of family. When equations, Xk=.oo88Xj+I.32 and Xk=2.38+.oo3Xj-.oo43Xl-.ooI35Xe, are utilized, the number of children per native white woman, ever married, 15-54, may be predicted from a given illiteracy rate. When the net regression equation is used, and Xi and Xe are held constant at their mean values, a zero illiteracy rate gives.4 children. The standard errors of estimate are Sk.jle =.18 8"Size of College Families," Journal of Heredity, October, 1924, p. 410. 9"Statistics of College and Non-College Women," Journal of the American Statistical Associatzon, VII (1900), 9. O00p. cit., chap. VII. ILLITERACY AND SIZE OF FAMILY 115 for the net regression equation and Ski =.3I for the gross regression equation. In measuring the relationship between illiteracy and size of family, we have again a biological reaction of man, affected by social factors. It is to be noted that the index of size of family utilized in this chapter, i.e., the number of children under fifteen years of age per ever married woman, 15 to 54 years of age, is a statistical approximation, in lieu of the practical impossibility of accurately determining the actual family size for all families in the United States. Size of family, as numerous studies in various countries have noted, is adversely affected by higher economic status. In addition, the number of children per family is lowered by increasing urbanization as various investigators have shown. The present chapter corroborates these findings. Moreover, the high negative correlations show the necessity of mathematically holding constant these two factors, which affect both illiteracy and size of family, if a truer picture of the relationship between these latter phenomena is to be obtained. The simple correlation between illiteracy and size of family is significant. When the effect of urbanization is controlled, the relationship is still important, and when the effect of economic status is controlled, there yet remains a tendency for illiterate mothers to have larger families than mothers of higher educational status. When both disturbing factors are controlled, the net correlation of +.36 continues to show an evident relationship between low educational status and large families. The regression equations once more offer a possibility of predicting that the size of family which is already quite small will continue to grow smaller with the increase in educational status. Moreover, the growing urbanization 116 ILLITERACY IN THE UNITED STATES rate and the apparent increase in the average income will hasten this shrinkage. Not a rosy picture, surely, for those who foresee the possible danger in this phenomenon! Lastly the importance of some socio-educational factor in analyzing the factors affecting the lowering size of family is borne out in this generalized study of the subject. CHAPTER XIV THE RELATION OF ILLITERACY TO MOBILITY TO OTHER STATES Present day society is characterized by a high degree of mobility. Probably one of the most important characteristics of contemporary culture is this intensive shifting of individuals from one locality to another.' The following chapter has for its object the measurement of the relation of one type of horizontal mobility to the factor of illiteracy. Interstate mobility is determined from the census which gives data pertaining to the percentage of various color and nationality classes who were born in a specified state, but who were, in I920, living in other states. For the purposes of this study, native whites of native parentage were utilized not only because the race and nationality factors are thus held constant, but also because interstate mobility of native whites of native parentage gives a more reliable statistical picture of actual mobility in the United States. The variable Xb is correlated with the variable, Xm, where Xb is the per cent of native whites of native parentage, ten years of age and over, who are illiterate,2 and Xm is the per cent of native whites of native parentage born in a specified state, but living in other states.3 The coefficient of correlation is -.40, the data being linear in character. This is interpreted as meaning that there is a low but significant negative relationship between illiteracy and 1Pitirim Sorokin, Social Mobility, p. 9. 2U. S. Census, 1920, vol. II, chap. XII, Table V. 'U. S. Census, 1920, vol. II, chap. V, Table XVI. 9 118 ILLITERACY IN THE UNITED STATES TABLE XXX DATA UTILIZED IN THE CORRELATION OF ILLITERACY AND MOBILITY TO OTHER STATES, NATIVE WHITES OF NATIVE PARENTAGE4 Percentage of Per Cent Percenfage of Percentage of Male Wage Illiterate, 10 Inhabitants Urbaniza- Earners in Years of Age Born in State tion 10 Years ManufacturState and Over and Living in of Age and ing 15 Years Other States Over of Age and Over Xb Xm Xi X. Maine............. 1.3 26. 4 31.7, 24.5 New Hampshire..... 0.6 37.2 49.1 35.6 Vermont............ 1. 37.5 26.2 22.4 Massachusetts...... 0. 3 20.1 90.3 36. 4 Rhode Island....... 0. 5 27.0 93. 9 42.6 Connecticut........ 0.4 24.2 57.9 45.4 New York.......... 0.6 19.8 67.1 23.5 New Jersey......... 0. 7 17. 6 68.4 35.0 Pennsylvania..... 0. 8 17.2 56. 3 30.3 Ohio.............. 1.0 22.5 55.0 29.4 Indiana............ 1.4 27.2 45.5 22.3 Illinois............ 1.1 30.6 52.1 19.9 Michigan........... 0.6 18.8 55.2 30.5 Wisconsin.......... 0. 5 24.9 44.1 28. 1 Minnesota........... 0.4 26.2 47.0 11.1 Iowa.............. 0. 5 37.0 38.3 7.9 Missouri............ 2.2 33.2 38.9 12.1 North Dakota...... 0. 3 32.6 20.2 1.9 South Dakota....... 0.3 32.7 21.3 2.5 Nebraska........... 0.4 33.9 32.6 6.8 Kansas............. 0.6 38.0 36.2 8.7 Delaware........... 2.0 31.2 45.9 30.3 Maryland.......... 2.0 20. 7 54. 6 20. 5 Virginia............ 6.1 22.7 28.6 13.2 West Virginia....... 4.8 18.1 25.0 15.3 North Carolina...... 8.2 14.3 19.6 15.4 South Carolina...... 6.6 15.3 22.1 12.4 ILLITERACY AND MOBILITY 119 TABLE XXX-Continued Percentage of Per Cent Percentage of Percentage of Male Wage Illiterate, 10 Inhabitants Urbaniza- Earners in Years of Age Born in State tion 10 Years ManufacturState andOver and Living in of Age and ing 15 Years Other States and Over of Age and Over Xb, Xm Xi X, Georgia............ 5.5 18.0 27.5 11.2 Florida............. 3. 1 ' 13.7 32.4 20. 6 Kentucky........... 7.3 25.7 22.2 7.4 Tennessee........... 7.4 26.6 23.6 10.2 Alabama........... 6.4 22.6 21.6 13.6 Mississippi........... 3.6 27.1 16. 6 10.0 Arkansas........... 4. 6 30.5 17.3 8.6 Louisiana........... 11.4 12.9 36.0 15.3 Oklahoma.......... 2.4 23.1 28.1 4.1 Texas.............. 2.2 16.2 32.7 6.3 Montana............ 0. 3 32.1 31.4 7.9 Idaho............... 0.3 29.4 28.9 8.6 Wyoming.......... 0.4 39.8 28.7 8.3 Colorado............ 1.7 34.5 48.6 9.1 New Mexico........ 11.9 21.8 18.8 4.6 Arizona............. 1.3 30.7 40.4 6.6 Utah.............. 0.3 22.6 46.9 11. 1 Nevada............ 0.4 53.9 23.6 8.4 Washington.......... 0.3 23.6 54.2 23.0 Oregon.............. 0.4 27.0 47.3 17. 4 California............ 0.4 11.6 68.1 14. 7 Mean.............. 2.4 26.0 40.0 16.9 4Sources: Xb-Fourteenth Census of the United States, vol. II, chap. XII,Table V; Xm-Ibid., chap. V, Table XVI; Xi-Ibid., computed from chap. I, Table XX, and chap. II, Table XIII; Xn-Ibid., computed from chap. III, Table XIII, and vol. VIII, Table XV. 120 ILLITERACY IN THE UNITED STATES interstate mobility. The significance of the result is more apparent when it is pointed out that certain factors, important in themselves, tend to obscure the actual correlation. The urbanization of a state would tend to affect the correlation, due to the fact that where a state is highly urbanized, this acts as an attracting factor to that state's ruralpopulation, and hence acts as a partial deterrent, it is believed, to migration to other states; moreover, the presence of a nearby urbanized state would act as an attracting factor to inhabitants of other states. Therefore, it is a factor which should be held constant, and this should raisenot lower-the net correlation over the gross correlation. A second factor of importance, related to the factor just discussed, yet different from it, is the amount of industry within a state. Where a state is highly industrialized, it acts as a deterrent to migration elsewhere, it is contended, as well as an attracting stimulus-situation to nearby less industrialized states. Partial correlation is therefore resorted to, as the next logical step in the analysis of illiteracy in its relation to mobility to other states. The notation employed is as follows: Xb = percentage of illiteracy, native white of native parentage, Io years of age and over. X= percentage of native whites of native parentage born in a specified state and living in other states. Xi= percentage of urbanization of native whites of native parentage, 10 years of age and over. Xn = percentage of male wage earners in manufacturing, 15 years of age and over. The last variable was utilized as an index of industrialization. The percentage of actual wage earners was considered a more fitting index than the amount of capital em ILLITERACY AND MOBILITY 121 ployed in industry because it is after all the opportunity afforded men that is the factor necessary to hold constant.5 Table XXXI presents the correlations of the zero, first, and second orders. TABLE XXXI GROSS AND NET CORRELATIONS r- = -.40 rnb = -.-.51 rh = -.49 rmi.n =-.06. rbn=-.28 rmn.b= -.40 rmi= -.22 rtmn. = -.11 rmn= -.24 rin.b= +.76 rin= +.78 rn.m = +.76 r&m. = -.59 bm.in = -.58 rbm.n= -.50 i.mn= -.55 rbi.m=-.64 rbn.mi= +.14 fbi.n = -.45 rmi.bn =-.36 rbn.m = -.42 mn.bi -.006 nbn.i +.1 8 rin.bm= +.71 When the percentage of urbanization is held constant, the gross correlation of -.40 becomes a correlation of -.59, (rbm.i= -.59) a result to be expected, if the preceding argument holds. When the percentage of male wage earners in manufacturing is held constant, the result is a correlation of -.50 (rbm.n=-.50) which is also to be expected. When both percentage of urbanization and percentage of male wage earners in manufacturing are held constant, rbm.in -.58. This correlation emphasizes the important negative relationship existing between illiteracy and mobility to other states since two of the strongest affecting factors are controlled. Further analysis of the net correla5Computed from data obtained from U. S. Census, 1920, vol. VII. Table XV, and vol. II, chap. III, Table XIII. 122 ILLITERACY IN THE UNITED STATES tions reveals the fact that this partial correlation is higher than any other combination of factors except percentage of urbanization with percentage of male wage earners in manufacturing (rin.bm= -.7I). It appears, therefore, that the degree of relationship between mobility to other states and illiteracy is higher than both the relationship between mobility to other states and percentage of urbanization and than the relationship between mobility to other states and the percentage of male wage earners in manufacturing, I5 years of age and over. Hornell Hart's analysis of data secured for the state of Iowa also reveals this important negative relationship between illiteracy and mobility.6 Robert E. Park points out that education is one of the factors which has operated to vastly increase the mobility of modern peoples.7 In analyzing English labor conditions of the nineteenth century, Redford found that "Peasants, especially uneducated peasants, are immobile except under the spur of extreme necessity. They do not migrate whenever it is to their economic interest, but only when they must."8 Since the primary interest is the prediction of mobility to other states on the basis of a given illiteracy rate, the following equations may be utilized. Xm = -.IO9Xb+28.75 Xm= 40. 55 - 1.849Xb -.248Xi+.o43 5Xn The corresponding standard errors of estimate areSmb = 7.6I and S.bin =6.53. When the net regression equation is utilized and when Xi and X, are held constant at their respec6"Selective Migration as a Factor in Child Welfare in the United States with Especial Reference to Iowa," University of Iowa Studies in Child Welfare, vol. I, no. 7, p. 56. 7"The City; Suggestions for the Investigation of Human Behavior in the City Environment," American Journal of Sociology, XX (1915), 589. 8Arthur Redford, Labour Migration in England, 1800-1850, p. 82. ILLITERACY AND MOBILITY 123' tive means, the substitution of an illiteracy rate of zero gives 30.55 as the predicted percentage of mobility to other states. By' way'of summary, it may be pointed out that the present day shifting of individuals from one state to another is correlated with various factors. The present chapter has shown that migration'to other states is partially hindered when a particular state offers certain attracting alternatives. For purposes of this study, degree of urbanization and individual opportunity have been selected as factors' which act as deterring agents to migration to other states. Ir order to obtain a true picture of the possible relationship existing between illiteracy and mobility, to other states, it would be necessary, first, to eliminate race and foreign birth factors, and, second, to control the factors of urbanization and industrial opportunity. Race and foreign birth may be statistically eliminated by utilizing the sub-classifications of the Census Bureau. The urbanization and industrial factors are controlled through the utilization of the method of partial correlation. The simple coefficient of correlation between illiteracy and mobility to other states is a low but significant one. When either percentage of urbanization or the percentage of male wage earners engaged in manufacturing (which is utilized as the index of industrialization) is held constant, the correlation between the primary factors of illiteracy and mobility to other states rises, bearing out the hypothesis that the deterring factors disturb the relationship between illiteracy and mobility. When both affecting factors are controlled, the net relationship is important enough to support the hypothesis on which the present chapter is based, namely, that illiterates, or persons of low educational status, tend to be more stagnant than persons of higher 124 ILLITERACY IN THE UNITED STATES educational status. There are various factors-which need not be entered into here-which tend to restrict more severely the average horizon of the person of low educational status, causing him to remain in a more fixed locality, often at the expense of finding greater opportunity, especially for one of his status, elsewhere. Regression equations enable one to predict that as illiteracy decreases and educational status rises, mobility will tend to increase, particularly from states with lesser opportunity for obtaining the primary needs of life. It would appear that the analysis of factors leading to greater or less mobility must take into consideration some index of socio-educational status, if the analysis is to be a complete one. CHAPTER XV THE RELATION OF ILLITERACY TO SUICIDE Suicide is a form of adjustment which results in death. From the socio-psychological standpoint, suicide is behavioristically a reaction to the complexity of social stimuli. From the standpoint of the present social group, suicide is an undesirable phenomenon, striking as it does at the all-important (from the standpoint of the group) social organization. The hypothesis advanced in this chapter is that there is an inverse relationship of illiteracy with suicide.1 Since suicide data are not presented by the census according to race and nationality, data for all classes were utilized in the present investigation. The percentage of illiteracy for all classes ten years of age and over was correlated with the number of suicides per 00oo,ooo population. Since suicides do not occur in large numbers the suicides for I9I9, 1920, and 19212 were averaged for each of the thirty-four registration states and used as the basis of computation. The gross correlation for these two variables is -.74, which is indicative of a high negative relationship. The data were previously tested and found to be linear. It was necessary to refine the data to a greater degree by the utilization of partial correlations. The percentage 1This does not imply a causal relationship. In fact, the concept of "cause" is foreign to any discussion in this study. Factorization, of which the end result is but a function, and that within the limits of the data, is always implied. 2Mortality Statistics, 1919, 1920, 1921, Table VIII. 126 ILLITERACY IN THE UNITED STATES TABLE XXXII DATA UTILIZED IN THE CORRELATION OF ILLITERACY AND SUICIDE, ALL CLASSES3 Percentage No.of Percentage Percentage Percentage of Illit- of Urban- Not ln Not Itc Suicides* Married eracy, 10 ization St Native 10 Per State 15 White of State Years of 100000 10 Years White of Age and Popul n of Age and Years o Native Over Over and Parentage Over Xf Xo Xp Xq XI California........ 3.3 24.0 69.2 42.6 51.0 Colorado........ 3.2 14.9 50.9 39.8 35.8 Connecticut...... 6.2 13.3 67.7 41.4 67.5 Delaware........ 5.9 9.1 54.6 39.4 37.3 Florida.......... 9.6 7.3 38.6 37.6 45.0 Illinois'........... 3.4 13. 5 68.6 40. 3 52. 7 Indiana......... 2.2 12.8 51.4 36.9 20.5 Kansas.......... 1.6 11.5 36.2 37.7 26.0 Kentucky....... 8.4 7.3 28.6 38.2 15.6 Louisiana........ 21.9 5.6 37.5 40.6 47.6 Maine........... 3.3 12.6 39.6 40.4 35.4 Maryland....... 5.6 10.8 61.2 41.9 38.4 Massachusetts... 4.7 11.2 94.7 44.7 68.1 Michigan......... 3.0 10. 6 61.9 37.6 54.5 Minnesota........ 1.8 12.5 45.5 44.5 65.3 Mississippi...... 17.2 3.2 14.6 37. 7 53.8 Missouri........ 3.0 12.7 48.7 39.3 25.5 Montana........ 2.3 15.9 33.0 40.2 49.8 Nebraska......... 1.4 12.3 32.8 39.7 41.6 New Hampshire.. 4.4 12.0 62.8 42.1 49.1 New Jersey...... 5.1 12.7 78.3 39.8 61.6 New York....... 5.1 12.6 82.6 42.4 64.7 North Carolina.. 13.1 3.8 20.8 39.4 31.0 Ohio............ 2.8 11.9 64. 6 38.5 36.3 Oregon...........1.5 14.1 51.4 39.4 36.5 ILLITERACY AND SUICIDE 127 TABLE XXXII-Continued Percentage Percentage No.of Percentage Percentage of Illit- Suicides* of Urban- Married Not eracy, 10 er ization ate 15 Native Per State, 15 State Years of 100,000 10 Years Years of White of Age and of Age and Native Population of Ag e and Native Over Over Over Parentage Xf X.o P X X Pennsylvania... 4.6 10.0 65.4 40.3 45.5 Rhode Island.... 6. 5 9. 7 97.4 44.2 71. 3 South Carolina... 18.1 3.9 19.2 39.8 52.5 Tennessee....... 10.3 6.3 28.3 37.9 21.6 Utah............ 1.9 8.8 50.2 39.2 45.3 Vermont........ 3.0 13.0 31.6 40.1 35.2 Virginia......... 11.2 7.0 31.3 41.3 33.5 Washington...... 1.7 17.6 56 7 40.4 47.5 Wisconsin....... 2.4 11.7 48.4 42.0 59.9 Mean.......... 5.9 11.1 50.7 40.2 44.8 *Average number of suicides for 1919, 1920, 1921. of urbanization4 for all classes, ten years of age and over, was adopted as the third variable. Since it is a known fact that suicide rates are higher among the non-married5 than among the married, the percentage not included in the married state, fifteen years of age and over,6 was taken as aSources: Xf-Fourteenth Census of the United States, vol. II, chap. XII, Table XII; X —Mortality Statistics, 1919, 1920, 1921, Table VIII; Xp-Fourteenth Census of the United States, vol. II, chap. XII, computed from Tables XII and XXII; Xq-Ibid., chap. IV, computed from Table XI; X,-Ibid., chap. I, computed from Table VII. 4Ibid., chap. XII, computed from Tables XII and XXII. 5Single, widowed, and divorced. 6Fourteenth Census, vol. II, chap. IV, computed from Table XI. 128 ILLITERACY IN THE UNITED STATES the next variable. Since suicide rates differ significantly according to race and nationality groupings,7 this factor was controlled by taking as a fifth variable, the percentage not native white of native parentage.8 The system of notation employed is as follows: f=Percentage of illiteracy for all classes 10 years of age and over. o= Suicides per Ioo,ooo population, all classes. p=Percentage of urbanization of all classes, Io years of age and over. q=Percentage not in the married state, all classes, 15 years of age and over. s = Percentage not native white of native parentage, all ages. The correlations of the zero, first, second and third orders are presented in the following table: TABLE XXXIII GROSS AND NET CORRELATIONS rf o=-.74 rfp.o=-.13 roq.p= +.07 r/p= -.43 r/p.q= —.44 ro.,= +.37 r/f= -.14 r/p.8=-.52 ro./ = +.24 ri,= +.006 rfq.o= +.19 ro,.p=-.15 rop= +.47 rlq.p= +.19 r8.,=-.19 rog= +.35 r/o.t= -.23 rpq.f = +.65 r0,,=+.16 rf.o= +.l18 rp2.o=+.57 rpq=+.64 rf8.p=+.33 rpq.,=+.39 rp = +.56 rf.q= +.18 rpe.f= +.63 rq,= +.77 rop. = +.26 rp,.= +.56 rfo.p=-.67 rop.q= +.34 rp,.q= +.15 r/o.q= -.74 rop.= +.47 rqs.f= +.78 ro.,= —.75 rq.f= +'37 rq,.= +.77 7For example, Negroes have low rates; Japanese have high rates; North Europeans generally have high rates, etc. See A. D. Frenay, The Suicide Problem in the United States, chap. XI. 8Fourteenth Census, vol. II, chap. I, computed from Table VII. ILLITERACY AND SUICIDE 129 TABLE XXXIII-Continued rqs.p= +.61 rop./ = +.15 rqa.lO= +.77 rfo.pq=-.70 top., = +.38 r2,.p = +.63 =ro.vp= -.67 oq.fp = +.28 r,.op = +.66 rfo.qs = —.73 toq., = +.3 1 rfp.oq= -.30 ro.p= +.21 rfo.pqs= -.68 rfp.oa= -.32 ros.fp= +.10 rfp.oq = -.32 rfp.qs= -.48 ros.f = -.09 r/q.op,= +.16 rfq.op = +.32 ro.pq =-.25 rf.opq = +.14 rq.os = +.08 rp.fo = +.61 rop.fq= +.05 rfq.p8= -.02 rpq.fs= +.32 roq.fp,= +.28 rf.op = +.31 rpq.os= +.26 ro8.fp = -.10 rlf.0o= +.06 rp,..fo= +.60 rpq.fos= +.30 rfs.pq= +.28 rp,.q= +.26 rpa.foq= +.26 rop. fq +.03 rp.og = +.24 rqs.fop = +.63 The gross correlation between illiteracy and suicide is rfo= -.74. When the factor of urbanization is controlled, rfO.p = -.67. When the factor of percentage not in the married state is held constant, rfo.q= -.74. Finally, when the percentage not native white of native parentage is controlled, rfo.8 = -.75. In the correlations of the second order, rfo.pq= -.70, rfo.ps= -.67, and rfo.qs= -.73. That there is a high negative degree of relationship between illiteracy and suicide appears to be substantiated by each of these correlations. Furthermore, in his analysis of European data, Morselli also shows the negative relationship between illiteracy and suicide.9 Since there is an inverse relationship between illiteracy and suicide, the probable suicide rate may be predicted from the following equations when the illiteracy rate falls to an approximate zero. X0= -5.885Xf+45.82, and X = - 3.o5 -.604Xf +.0225Xp +.4743 Xq- o5 57Xs. 9Suicide, chap. IV. Suicide, of course, refers to the suicide rate per 100,000 population, as defined on page 129, above. 130 ILLITERACY IN THE UNITED STATES The standard errors of estimate are Sof = 2.77 and So.fpqs = 1.26 respectively. Substituting the mean values of Xp, Xq, and Xs and reducing Xf (illiteracy) to zero, Xo or the suicide rate becomes 14.66 per IOO,OOO. In other words, if conditions remained the same as utilized here and if the illiteracy rate dropped to zero, the chances are that in 68 out of Ico cases, the suicide rate would fall within the limits of I4.66 = I.26. In summing up, it may be pointed out that suicide is a form of-from the standpoint of the group-anti-social behavior. Hence the mores operate more or less in an unconscious endeavor to prevent the taking of one's own life. Does low educational status, as reflected in illiteracy, reduce suicide or is the opposite true? This is the question raised in the present chapter and an answer is attempted. When the relationship of illiteracy to suicide is measured, a high negative correlation,-.74,, is obtained. Apparently illiterates do not commit suicide, as compared with persons of higher educational status. It is claimed, however, that suicides are committed in urban areas, and since a large percentage of illiterates live in rural areas, the above correlation is affected. The method of partial correlation enables one to control the factor of urbanization, giving a resulting correlation of -.67, which is still significant. There is another factor of importance to be taken into consideration. Suicide rates are higher among the nonmarried (single, widowed, or divorced) than among the married, and hence it is desirable to control this differential factor. When the percentage not included in the married state, fifteen years of age and over, is held constant, the correlation first obtained, -.74, is unchanged. Again suicide rates in the United States differ significantly according to the races, and also according to coun ILLITERACY AND SUICIDE 131 try of birth. These latter two factors may be controlled at the same time by holding constant mathematically the percentage not native white of native parentage. When this is done, the resulting correlation remains practically unchanged. Various correlations of the, higher orders are presented in the tables for purposes of completeness, but due to the'small number of cases, their validity is questionable. The net coefficient of correlation between illiteracy and suicide, with all the disturbing factors mentioned above controlled, however, is consistent with the results arrived at previously, and is found to be -.68, an apparently significant correlation. It would appear, then, that the illiterate, as compared with persons of higher educational status, do not commit suicide to an important degree. When the regression equations are utilized, the suicide rate may be predicted as rising as educational status rises. To what degree this 'is due to the rise of educational status, and to what degree it is due to the increased complexity of stimuli which confront the person of higher educational status is an important study in itself. It may be pointed out that from the results obtained it is evident that the inverse relationship between'illiter'acy and suicide is so high that no investigator of the 'social factors affecting suicide can neglect the phenomenon of illiteracy or some socio-educational equivalent. CHAPTER XVI THE RELATION OF ILLITERACY TO URBANIZATION AND TO SCHOOL SYSTEMS The present chapter has for its objective the measurement of the relationship of illiteracy to urbanization and to school systems. Figures have been given' which show the higher rate of illiteracy in rural than in urban areas for the United States in both I9Io and 1920. The correlations secured in each chapter of Part II indicate a relationship between illiteracy and urbanization as well as a relationship between the other primary factor being studied and urbanization. Is literacy, then, strictly speaking, a function of urbanization? Or is it rather that urbanization possibly provides the opportunity to society of maintaining adequate school systems which in turn operate to reduce the rate of illiteracy? Theoretically, illiteracy may be reduced to an approximate zero because the human organism is capable of attaining that degree of education which would classify him as literate according to the census terminology. The exceptions to this include the relatively small percentage of those who are idiots and low grade imbeciles, as well perhaps as a few other physical defectives. Insanity, almost without exception, occurs sufficiently late in life so that if a person were capable of becoming literate he would already have become so provided there were adequate educational opportunity. Germany, Holland, and Switzerland, for example, have practically negligible rates of illit1See above, Chap. V. ILLITERACY AND URBANIZATION 133 TABLE XXXIV DATA UTILIZED IN THE CORRELATION OF ILLITERACY AND URBANIZATION* Per Cent Percentage of Index Num- Index NumIlliteate 10 Urbaniza- bers for bers for ears of Ae 1tion 10 Years State School State School State and Over of Age and Systems Systems Over 1918 1890 Xb Xi Xt Xz Maine............. 1.3 31.7 47.4 29.9 New Hampshire..... 0.6 49.1 54.4 31.0 Vermont............ 1.1 26.2 51.5 30.2 Massachusetts...... 0.3 90.3 61.0 45.9 Rhode Island....... 0.5 93.9 56. 3 39.3 Connecticut.......... 0.4 57.9 59. 8 38.9 New York.......... 0.6 67.1 59.4 40.9 New Jersey......... 0.7 68.4 65.9 37. 5 Pennsylvania........ 0. 8 56.3 57.7 34. 7 Ohio............... 1.0 55.0 59.7 33.1 Indiana............. 1.4 45.5 58.8 29.8 Illinois.............. 1.1 52.1 56. 8 31.9 Michigan............ 0.6 55.2 60.4 31.9 Wisconsin...........0.5 44.1 51.3 31.0 Minnesota.......... 0.4 47.0 58.4 29.5 Iowa............... 0.5 38.3 61.9 31.0 Missouri............ 2.2 38.9 49.6 25.5 North Dakota...... 0.3 20.2 59.2 25.5 South Dakota....... 0.3 21.3 55.0 26.1 Nebraska........... 0.4 32.6 57.1 26.4 Kansas.............. 0.6 36.2 55.2 30. 6 Delaware........... 2.0 45.9 42.5 29.3 Maryland.......... 2.0 54. 6 43.2 33.3 Virginia............ 6.1 28.6 35.3 22.3 West Virginia....... 4.8 25.0 37.7 21.8 North Carolina...... 8.2 19.6 30. 6 17.8 South Carolina....... 6.6 22.1 29.4 12.5 Georgia............. 5.5 27.5 32.6 15.7 Florida............. 3.1 32.4 37.8 28.5 10 134 ILLITERACY IN THE UNITED STATES TABLE XXXIV-Continued Per Cent Percentage of Index Num- Index NumIlliterate 10 Urbaniza- bers for bers for Years of Age tion 10 Years State School State School State and Over of Age and Systems Systems Over 1918 1890 Xb Xi Xt Xz Kentucky........... 7. 3 22.2 35.0 23.4 Tennessee.......... 7.4 23.6 35.1 21.0 Alabama........... 6.4 21.6 30.6 18.2 Mississippi......... 3.6 16.6 30.0 21.9 Arkansas........... 4.6 17. 3 30. 3 20. 1 Louisiana.......... 11.4 36.0 33.9 18.4 Oklahoma.......... 2.4 28.1 44.4 20. 5 Texas.............. 2.2 32.7 41.1 23.2 Montana........... 0. 3 31.4 75.8 36.3 Idaho.............. 0.3 28.9 58.6 22.8 Wyoming........... 0.4 28.7 56.7 30. 3 Colorado........... 1.7 48.6 59. 2 37.8 New Mexico........ 11.9 18.8 53.0 10.0 Arizona............ 1.3 40.4 66. 2 32.8 Utah............... 0.3 46.9 61.4 28.6 Nevada............ 0.4 23.6 59.1 34.5 Washington......... 0.3 54.2 63.7 30. 8 Oregon............. 0.4 47. 3 57. 8 27.9 California.......... 0.4 68.1 71.2 43.8 M ean.............. 2.4 40.0 51.0 28.4 *Sources: Xb-Fourteenth Census of the United States, vol. II, chap. XII, Table V; Xi-Ibid., computed from chap. I, Table XX, and chap. II, Table XIII; Xr —Ayres, An Index Numberfor State School Systems, p. 37; X —Ibid., p. 31. The data on illiteracy and urbanization are for native whites of native parentage. The data on school systems are not given according to any race or "nationality" classification. ILLITERACY AND URBANIZATION 135 eracy.2 The methods by which this is ascertained, while not comparable with those of the United States, do give results which support the statements made above. The next logical step is a consideration of the adequacy of school systems as they exist in the United States today and as they have existed in past decades. The data presented by Leonard P. Ayres in "An Index Number for State School Systems" is utilized as forming the most adequate basis for analysis as was previously pointed out in Chapter VII. Before utilizing these educational data in the present chapter, the degree of gross relationship between illiteracy and urbanization was first investigated. The Pearsonian coefficient of correlation for illiteracy and urbanization, for all classes, was found to be -.36.3 Since the data were slightly non-linear in character, the index of correlation was computed on the basis of the following formula: Rh2 = a b( Y) + b(XY) + c(X2 Y) -Nc,2 Rho~, "= --- SP --- 2(Y2) - Nc2 Rho=-.39 which indicates that a second degree parabola affords a slightly better fit for the data than does a straight line. The data were then refined by utilizing only native whites of native parentage, ten years of age and over.4 The coefficient of correlation was found to be rbi= -.49. Since the data appeared on inspection to be somewhat nonlinear in character, the index of correlation was next computed on the basis of a second degree parabola, the formula given in connection with the discussion of the data for all classes being utilized. The result was Rhobi =.58. The non211iteracy and School Attendance in Canada (Dominion Bureau of Statistics, 1926), p. 30. 'See Appendix B for data utilized. 'Table XXXIV. 136 ILLITERACY IN THE UNITED STATES linear character of the data was further studied by means of an exponential curve. The formula, R2 log aly+log b2x. logy-Nc log 2 2(log y)2-Nc lo 2 was utilized. Proof of the fact that the data for illiteracy and urbanization, native white of native parentage, 10 years of age and over, fall along an exponential curve is given by the result, Rhobi=.74. Inasmuch as this is a gross correlation, it appears necessary to refine the data and determine whether the apparent high relationship between illiteracy and urbanization holds. L. P. Ayres's index numbers for school systems for the years 1918 and 1890 were utilized as the third and fourth variables. Since the relationship between each of these variables and illiteracy is linear, the four variables were utilized in the partial and multiple correlations on the assumption of general linearity. This could be done because when the non-linear relationship between urbanization and illiteracy, for native whites of native parentage, was tested for linearity where utilized in partial correlation, by computing the derivative and plotting against the dependent variable, the non-linear relationship was found not to be of significant effect. The size of the sample is also operative here. The notation employed was as follows: b = illiteracy rate for native whites of native parentage, 10 years of age and over. i= percentage of urbanization, native white of native parentage, 10 years of age and over. t=index numbers for state school systems, 1918. z=index numbers for state school systems, 1890. ILLITERACY AND URBANIZATION 137 TABLE XXXV GROSS AND NET CORRELATIONS r. = -.49 rit.b= +.32 rbt = -.72 rfi.z= -.08 rs= -.75 iz.b= +.74 rft= +.55 r. t = +.69 ri= +.79 ttz.b= +.42 frt= +.73 rt,.i= +.58 rbi. =-.16 7bi.tZ= +.25 rba.a= +.26 rbt.iz=-.38 rbt.i= -.62 rbz.i = -.50 rbt..= -.39 Fit.bz = +.03 ri.i = -.68 riz.bt= +.70,.t= -.41 rtz.bi= +.28 Since not only urbanization but also school systems are of primary interest here, the three gross correlations of rbi= -.49, rbt= -.72, and rb= -.75 may first be considered. The gross relationships of illiteracy and school systems for both I918 and I890 are higher than the gross relationship of illiteracy with urbanization. The correlation of illiteracy and school systems of I918, holding percentage of urbanization constant, is rbt. = -.62. The correlation of illiteracy and school systems of 1890, holding percentage of urbanization constant, is rb.i = -.68. As in the gross correlations of illiteracy with school systems for these years, a significant inverse correlation is found. Furthermore, in the correlations of both the zero and first orders, the data for school systems of I890 give a slightly higher correlation with illiteracy for 1920 than do the data for school systems of I918. In the partial correlations of the second order, rb.tz= +.25. When the effects of the school systems of I918 and 1890 are controlled, the relationship between illiteracy and urbanization becomes positive. This, however, is chiefly a 138 ILLITERACY IN THE UNITED STATES function of the index numbers for school systems in I890 when the United States was less highly urbanized. When the school systems in I918 alone are controlled rb. t = -.16, which appears to be the truer indication of the actual relationship at the present time. The coefficient of the multiple correlation is Rb.itz = 81I This is indicative of a high degree of relationship between the dependent variable, illiteracy, and the independent series in combination. In the preceding chapters of Part II, the regression equations have been presented at this point. Since the data for illiteracy and urbanization are exponential in character, similar predictions in this chapter would not be warranted. In fact, the regression equations were analyzed but the figures secured as a basis of prediction were obviously inaccurate. In speaking of the validity of such results where there is a departure from linearity, Mills says, "There would be no fallacy involved in the use of the equation under these conditions, but it would not furnish as good a basis for estimates as one which took account of the true relationship."5 In summary, the chapter pursues somewhat further the analysis of the already observed negative relationship between illiteracy and urbanization. There is, moreover, a significant relationship between illiteracy and the adequacy of state school systems for the various census years. The correlation remains important when percentage of urbanization is held constant. When the adequacy of school systems is controlled, however, the relationship between illiteracy and urbanization loses its significance. It would appear, therefore, that while illiteracy may be thought of as being negatively related to urbanization, more correctly 5Statistical Methods, p. 499. ILLITERACY AND URBANIZATION 139 speaking, the lower illiteracy rate is rather a function of the adequacy of urban school systems, in comparison with rural school systems. Finally, multiple correlation is resorted to in order to determine the effect of school systems and urbanization, in combination, upon educational status. The result, R=.8i, indicates the important relationship of the factors of urbanization and adequate school systems upon educational status. CHAPTER XVII CONCLUSIONS TO PART II Illiteracy in the United States is a steadily decreasing phenomenon. Nevertheless, in I920 there were in round numbers 4,900,000 persons in the United States over ten years of age who were illiterate. While the rate for the United States as a whole was six per cent, certain great groups of the population had a much higher rate. The human organism is a mechanism biologically equipped and culturally trained to make adjustments to life situations. Other factors being equal, it seems evident that the literate individual, on the average, is better equipped than the illiterate to adapt himself to existing situations or to problems as they arise. Mlodern culture is so dependent upon communication by means of written symbols that the illiterate individual is more definitely circumscribed in his ability to adjust to the surrounding culture than he would have been twenty or thirty or forty years ago. In Part II of the present study, an attempt has been made to analyze by statistical procedure the relationships which exist between illiteracy and certain selected social phenomena. First among these phenomena studied was birth-rate. Any factor which tends to either increase or decrease the birth-rate is a matter of significance. Various studies have been made in which the effect of education upon the birthrate has been pointed out. For example, the lowered birthrate among college graduates has practically become a truism. At least as significant an approach to the problem, CONCLUSIONS TO PART II 141 and certainly a more comprehensive one inasmuch as only small groups of college graduates have been studied, is an investigation of the relation of illiteracy to birth-rate for the country as a whole. Thus, when the two important affecting factors of urbanization and per capita current income were mathematically controlled, it was found that the net coefficient of correlation, indicating the degree of relationship between illiteracy and birth-rate, was r = +.49. This is indicative of the fact that high illiteracy rates and high birth-rates tend to go together. Furthermore, if other conditions remained the same and the illiteracy rate should fall to zero, it is possible to predict, by means of the regression equations, that there would be a significant decrease in the birth-rate for the country as a whole. After determining that the birth-rate is positively correlated with the illiteracy rate, the relation of this latter factor to deaths of infants under one year of age was next investigated. High infant mortality rates are to a great extent due to ignorance with regard to the proper care of children. It is obvious that the illiterate mother or father will be more ignorant in regard to modern methods of child care than the literate mother or father who has access to the great bulk of published material of a relatively scientific nature, dealing with the general question of proper care of infants. When the relationship existing between illiteracy and infant mortality is reduced to measurable terms, it is found that after urbanization and per capita current income are controlled, the net coefficient of correlation for the two factors of illiteracy and infant mortality is +.6i. In other words, high infant mortality rates are associated with high illiteracy rates. Further analysis of the existing data points to the fact that if other conditions were to remain the same while the illiteracy rate fell to an approximate zero, a significant decrease in the in 142 ILLITERACY IN THE UNITED STATES fant mortality rate of the United States would occur. Thus illiteracy is a factor which must be considered not only with regard to the birth-rate of the country but also with regard to the percentage of infants who do not survive the first year of life. The third social phenomenon to be investigated with regard to its relation to illiteracy was early age of marriage, considered in terms of the number of females, ever married, 15 to 19 years of age, for native whites of native parentage. The percentage of such early marriages in the population is a matter of definite social significance. "Our civilization is today so complex that the judgments and experience necessitated by family life are not always acquired early. Mariiage today also involves in many cases restrictions to opportunities and limitations on activities. Many occupations requiring long-time preparation in education or apprenticeship tend to discourage early marriage."l While these relationships are not specifically measured, nevertheless the above statements appear valid. It is possible, however, to measure the degree of relationship which exists between illiteracy and early age of marriage. When percentage of urbanization and per capita current income were held constant, the net coefficient of correlation between illiteracy and early age of marriage was found to be r= +.43. This means that there is a positive relationship between the percentage of illiteracy and the percentage of females between the ages of 15 and I9, native white of native parentage, who are married. Thus, it is to be expected (as the regression equations indicate) that if other conditions were to remain the same, a significant lowering of the illiteracy rate would be definitely connected with a decrease in the percentage of early marriage 'E. R. Groves and Wm. F. Ogburn, American Marriage and Family Relationships, p. 219. CONCLUSIONS TO PART II 143 among native white females of native parentage. The phenomenon next studied was that of size of family as defined in terms of the number of children under fifteen years of age for native white women, 15 to 54, ever married. This was believed to be a closer equivalent to the actual number of children per family than the census approximation. While size of family has been investigated from various standpoints, its relation to illiteracy has received but scant attention. According to the results of the present investigation, however, when the affecting factors of urbanization and per capita current income are mathematically controlled, the degree of relationship between illiteracy and size of family, as here considered, may be expressed in terms of the net coefficient of correlation as r= +.36. Thus, where illiteracy is high, the number of children per married woman tends to be relatively large. On the basis of the regression equations computed it is possible to predict that a decrease in the illiteracy rate, other conditions remaining the same, would lead to a decrease in the number of children per native white married woman within the specified age limits. A phenomenon which perhaps more definitely affects the redistribution of population than any of those hitherto considered is that of mobility. This factor has been analyzed here in terms of the percentage of native white persons of native parentage who were born in a specified state but were living in other states in I920. When the two factors of percentage of urbanization and percentage of male wage earners employed in manufacturing are held constant, the net coefficient of correlation for illiteracy and mobility to other states, for native whites of native parentage, becomes -.58. In other words, the illiterate individual tends to be less mobile than the literate person. Therefore, if the illiteracy rate of the United States should 144 ILLITERACY IN THE UNITED STATES approximate zero, and if other conditions should remain the same, the percentage of mobility to other states would apparently increase, with a resultant intensification of geographical redistribution of individuals in the United States. As in the case of mobility, suicide, the next phenomenon investigated, operates inversely with illiteracy. Various investigators have pointed out the fact that a relation seems to exist between suicide and degree of education. Thus suicide appears to be a form of adjustment which the literate tend to make rather than the illiterate. In measuring the relationship between illiteracy and suicide the gross coefficient of correlation for all classes was found to be r = -.74. Of the various social factors affecting suicide rates, the percentage of urbanization, marital status, and race and nationality were held constant in the partial correlations. When these-three variables were controlled, the net coefficient of correlation became r = -.68.This high negative figure emphasizes the fact that low illiteracy rates and high suicide rates are found together. Therefore, one may predict on the basis of the regression equation that if other conditions should remain the same and if the illiteracy rate should fall to zero, the suicide rate would increase. Although the factor of percentage of urbanization was taken into consideration in every social phenomenon studied, a more intensive analysis of this factor and of school systems was deemed advisable. From the unanalyzed data, it would appear that an important relationship exists between the illiteracy rate and percentage of urbanization. When the gross coefficient of correlation was computed for these two variables, for native whites of native parentage, it was found to be r = -.49. When index numbers for state school systems for i9i8 were held constant, the coefficient of partial correlation for illiteracy and percentage of ur CONCLUSIONS TO PART II 145 banization dropped to r= -.I6. The coefficient of partial correlation for illiteracy and index numbers for state school systems of I918, holding percentage of urbanization constant, was r= -.62. The coefficient of partial correlation for illiteracy and school systems of I890, again holding percentage of urbanization constant, was r= -.68. In other words, it would appear that the lower illiteracy rate for the urban areas of the United States as a whole is a function of the adequacy of urban school systems both at present and in past decades in comparison with rural school systems rather than a function of urbanization per se. In the analysis of the various societal phenomena which were studied, no cause and effect relationships were assumed. A process of factorization was followed and in each case, whether all classes, native whites, or native whites of native parentage were studied, at least two of the important factors affecting the phenomenon were controlled in the partial correlations, after the two primary factors had previously been refined as far as possible. Thus a measuring of relationships was the general objective and the reliability of these relationships was statistically determined by computation of the error.2 The phenomena studied in relation to illiteracy include birth-rate, infant mortality, early age of marriage, size of family, mobility and suicide. As a general conclusion, it may be stated that the rate of illiteracy is a factor which is related to each of these societal phenomena and which has, in general, not been adequately considered in previous investigations. It is believed, on the basis of the results obtained, that no comprehensive analysis of any one of these subjects can omit the important, measurable factor of illiteracy or some socio-educational equivalent. Within the 2A table of the errors of coefficients of correlation is included in Appendix A. 146 ILLITERACY IN THE UNITED STATES limits of the present investigation, the importance of education as related to various types of social data has been demonstrated. APPENDIXES APPENDIX A METHODOLOGY Appendix A contains a brief statement of the statistical techniques employed in the analysis of illiteracy and certain related factors. To some extent, the classificatory procedure utilized by the census bureau provided a method of refining data. Part I of the present study utilizes this in the analysis of illiteracy data, according to sex, age,; race and nationality, urban and rural environment. It may be added that this method of refining data sometimes obviates the use of partial correlation. In Chapter II, straight line trends were put through the data for the United States as a whole and for each state for the period, 1870 to I920.1 From observation of the plotted data, it was evident that the trends of illiteracy over, the period studied were in general linear in character. A straight line was fitted to the six points corresponding to theisix decades by the method of least squares, the following equations being used: 2(y) =na+bZ(x) - (xy) = a2(x) +bb(x2) Solving these two equations, the most probable relationship between the variables studied was obtained.2, In some cases straight lines were also computed for shorter periods while the data for a few states appeared somewhat curvilinear. In these latter cases a second degree parabola was utilized, but it was found that it afforded no better fit to the data. 2For a discussion of the theory on which the method is based see Frederick C. Mills's Statistical Methods, pp. 273-78. Also, see C. G. Dittmer's Introduction to Social Statistics, pp. 149-52. 11 150 ILLITERACY IN THE UNITED STATES The statistical tool known as the method of standard population was utilized in Chapter VIII. At that point a discussion of the methods employed was given in some detail.3 Utilization of this method enables the investigator to control the important factor of age distribution. The methodology employed in Part II consisted in the utilization of certain additional statistical tools. The results obtained with these techniques were discussed as they were presented. As a preliminary investigation, the Pearsonian coefficient of correlation was computed for illiteracy and each of the social phenomena studied4 as an abstract measure of the degree of relationship existing between the two variables. The value of r, the symbol for the coefficient of correlation, fluctuates between o and = I. A correlation of zero indicates no relationship while a correlation of one indicates a perfect relationship which may be either positive or negative.5 Owing to the complexity of social data, a valid, perfect correlation is practically an impossibility. In general, the higher the correlation and the lower the error, the more significant does the former become. The formula utilized in computing the Pearsonian coefficient of correlation was 3For further discussion, consult G. C. Whipple's Vital Statistics, pp. 292-303. 4Divorce was investigated but proved unsatisfactory for the present study owing to the lack of standardization in divorce laws. Insanity also appeared unsatisfactory as the rates are so dependent on the states' provisions for the care of persons with mental diseases. Statistics on crime are largely dependent on the various state laws and criminal procedures so that they, also, could not be utilized. In all three cases the Pearsonian coefficients of correlation were computed but the errors were so large that the correlations were not significant. 6For a clear discussion of the meaning and methodology of the Pearsonian coefficient, see Edmund E. Day's Statistical Analysis, pp. 193-99. A more extended discussion of linear correlation is found in G. Udny Yule's An Introduction to the Theory of Statistics, chap. IX. METHODOLOGY 151 (x'y') (X -- CX C _ -cc \ N cx N cy2 In certain instances, however, the relationship between the two variables was found, upon plotting, to be nonlinear in character. The measure of non-linear relationship is called the index of correlation and is represented by the symbol Rho. The index of correlation6 was computed for curves that were parabolic in nature and also for exponential curves. The formula utilized when a second degree parabola was fitted to the data was a 2(X) +b (X Y) + c (X2Y) -Nc,2 Rho2yz =; (Y2) - Nc 2 When the data were exponential in character, logarithms were utilized in the following formula for the computation of the index of correlation: log a2 y + log b; x. log y- Nclog v2 Rho2 ( = Rhovx,:(log y)2-Nclg,2 In computing the coefficients of partial and multiple correlation, two methods were followed. When five variables were utilized, it was considered perhaps more economical to step the coefficients of correlation up from the zero order, to the first order, and so on.7 For example, after computing the r's of the zero order, the r's of the first order were computed from the formula, 6See Frederick C. Mills's op. cit., pp. 436-41. 7For an analysis of the theory and methodology, consult G. Udny Yule's An Introduction to the Theory of Statistics, chap. XV. Also, see Frederick C. Mills's Statistical Methods, pp. 507-12. 152 ILLITERACY IN THE UNITED STATES r'12- r13 ~ r23 r12.3 (I -r213)Y2 (I -r223) Y2 This formula was further expanded for the computation of the coefficients of correlation of the second and third orders. The coefficients of regression were computed from equations of the form b12.345=r12.345 S-345 The standard S2.345 error of estimate for five variables was obtained by means of the formula, S21.2345= (I -r212) ( -r213.2) (I -214.23) ( -r215.234) In computing the coefficient of multiple correlation denoted by R, which measures the degree of relationship between the single dependent variable and the series of independent variables, in combination, the equation R2 21.2346 R21.2345 = I. —2 o12 was utilized. The coefficients of partial and multiple correlation were, in general, computed for only four variables, however. In those cases the first method of computingt he coefficients given by Frederick C. Mills in his Statistical Methods8 was followed. The coefficients of net correlation were computed on a basis of the coefficients of regression. For example, r12.34= V b12.34. b21.34 The b's were previously computed from the normal equations according to the Doolittle Method of solution.9 The estimating or regression equation took the form, 8Chapter XIV. This chapter includes a discussion of the various tools of analysis utilized. 9Mills, op. cit., pp. 577-81. M..ETHODOLOGY 153 Xl = a+- b12.34 X2 +b13.24 X3+^b14.23 X4 The reliability of estimates based on this equation was determined by the computation of the standard error of estimate from the formula., -_. S21.24 = 12- b12.34 P12 - b-124 p13 - b14.23 P14 The coefficient of multiple correlation which is dependent on the relation between S and a was computed from the formula, S21.234 R21.234 = I ---a12..I The standard error of the correlation coefficientl0 is I - r2 computed according to the formula, oa= i N This formula holds for the coefficient of multiple correlation and coefficients of partial correlation as well as for the simple coefficient of correlation. The following table gives the errors for the various numbers of states employed: According to W. F. Ogburn,1l the coefficient of correlation should be at least three times as large as its error bebefore one can be sure it is significant. Other authorities12 make comparable statements. Some prefer to state the error in terms of the probable error or.6745 times the error as computed above. At this point, some further explanation as to the theory of correlation should perhaps be added. Day states this concisely. "By correlation is meant, in brief, a definite '0Yule, op. cit., p. 352; also, Mills, op. cit., p. 556. Note: Scatter diagrams were prepared for all intercorrelations as a test for linearity. The single exception to linearity or approximate linearity is dealt with on page 136 above. "Groves and Ogburn, American Marriage and Family Relationships, p. 474. 12See Harry Jerome, Statistical Method, p. 285. 154 ILLITERACY IN THE UNITED STATES tendency for two or more variables to vary together. The variables may move in the same or in opposite directions, but if they are correlated they are never indifferent to one another, they are either mutually attractive or mutually repellent. Correlation involves a one-to-one correspondError When Error When Error When Error When r N=48 N=34 N=33 N=21 0.05 0.144 0.171 0.174 0.218 0.10 0.143 0.170 0.172 0.216 0.15 0.141 0.168 0.170 0.213 0.20 0.139 0.165 0.167 0.209 0.25 0. 135 0.161 0.163 0.205 0.30 0.131 0.156 0.158 0.199 0.35 0.127 0.150 0.153 0.191 0.40 0.121 0.144 0.146 0.183 0.45 0.115 0.137 0.139 0.174 0.50 0.108 0.129 0.131 0.164 0.55 0.101 0.120 0.121 0.152 0.60 0.092 0.110 0.111 0.140 0.65 0.083 0.099 0.101 0.126 0.70 0.074 0.087 0.089 0.111 0. 75 0.063 0.075 0.076 0.095 0.80 0.052 0.062 0.063 0.079 0.85 0.040 0.048 0.048 0.061 0.90 0.027 0.033 0.033 0.041 0.95 0.014 0.017 0.017 0.021 ence between the paired variables."'3 The essential point is the mutual relationship between two variable phenomena. The closeness of this relationship determines the size of the coefficient of correlation, r. Yule, in discussing the coefficient of correlation, says, "The constant r is of very great importance. It is evidently a pure number and its magnitude is unaffected by the scales in which x and y are l"Edmund E. Day, Statistical Analysis, pp. 188-89. METHODOLOGY 155 measured, for these scales will affect the numerator and denominator to the same extent."" Multiple and partial correlation methods enable the investigator to carry his analysis much farther than in simple correlation alone and to more nearly approach the work of the exact scientist who controls his various factors. "The coefficient of multiple correlation is an index of the degree of relationship between a single dependent variable and a number of independent variables, in combination. It measures the degree to which variations in the dependent variable are related to the combined action of the other factors. Its significance may be clearer if all the independent variables are looked upon as constituting a single independent series. The coefficient is then seen to be a measure of the relationship between the dependent variable and the independent series, which is precisely what the coefficient of correlation is in the simpler case of two variables."15 Partial or net correlation signifies the degree of relationship between two variables when other specified factors are held constant. That is, the investigator can theoretically determine, within the limits of the error, the degree of association between two variables when other affecting variables (most often one or two) are abstracted. The importance of this statistical tool is perhaps obvious. The caution should be included, however, that where the number of cases is small, as in the present study, too much reliance upon the mathematical values of the partial correlations is both unwarranted and dangerous. Because of the recognition of this fact, the interpretations in the body of the study must be regarded as empirical, inferential probabilities, based upon a thorough knowledge of the data. 14G. Udny Yule, An Introduction to the Theory of Statistics, p. 173. 15Mills, op. cit., pp. 497-98. 156 ILLITERACY IN THE UNITED STATES It is admittedly hazardous to lay down any specific rules for the interpretation of a coefficient of correlation as the type of-data studied and the number of cases involved must be considered as well as the actual size of the coefficient itself. "Furthermore, it is more hazardous to attempt to appraise the partial dependence of one variable on each of several others than to estimate the degree of relationship between two variables."' In addition, the inferences drawn from the use of such methodology must be carefully stated. All careful statisticians emphasize the fact that generalizations must not only be consistent with the data but also that they must be kept within the limits of the data presented. Furthermore, the present study does not pretend to give a complete description of the data. Only those aspects necessary for the development of the hypotheses presented were utilized. In the problem, therefore, attempts at interpretation were definitely limited to such generalizations as the data clearly justified. Other possible interpretations were rigidly excluded as being inconsistent with the purpose of the investigation. Finally, for the technical justification for correlating ratios and unweighted averages, reference may be made to Frank A. Ross's monograph, "School Attendance in 1920".17 16W. L. Crum, and A. C. Patton, An Introduction to the Methods of Economic Statistics, p. 263. 17U. S. Bureau of the Census, Monograph V, pp. 210-11. See also, M. R. Neifield, "A Study of Spurious Correlation," Journal of the American Statistical Association, XXII, 331-38, and G. Udny Yule, "On the Interpretation of Correlation, between Indices or Ratios," Journal of the Royal Statistical Society, vol. LXXIII (1910), pp. 644-47. APPENDIX B Appendix B contains supplementary tables, the data having been utilized only in simple correlations. All other data are included in the chapters of which they form the basis..:. I I, I I I I 7 t 158 ILLITERACY IN THE UNITED STATES TABLE XXXVI DATA FOR CORRELATION OF ILLITERACY AND BIRTII-RATE, FOR NATIVE WHITE FEMALES, 15-441 Illiteracy Rate* Legitimate2 Births** State Per 1,000 Married Females California..................... 3. 4 119. 4 Connecticut......................3. 5 153. 3 Indiana...................... 5.1 150. 1 Kansas...................... 3. 2 157.3 Kentucky.................... 43.5 188.9 Maryland..................... 9. 1 170.8 Massachusetts.....................3.0 169. 4 Michigan.................... 3. 7 167.4 Minnesota................... 2.9 189. 3 Nebraska..................... 2.6 170. 1 New Hampshire.............. 5.6 167. 1 New York.................... 3.4 146.8 North Carolina....................6.0 231.1 Ohio........................ 3. 8 140.9 Oregon....................... 2.1 125.8 Pennsylvania................... 4. 5 171. 6 South Carolina....................68. 6 205. 8 Vermont.................... 7. 2 157.9 Virginia..................... 36. 4 230. 5 Washington.................... 1.6 131.6 Wisconsin.................... 3. 8 172. 8 'Sources: *Fourteenth Census of The United States, vol. II, chap. XII, computed from Table XII. **Birth Statistics for the Birth Registration Area of the United States, 1920, Table V, and Fourteenth Census, vol. II, chap. IV, Table XI. 2Unknown legitimacy is tabulated as legitimate while the birth certificates of California, Massachusetts, New Hampshire and Vermont do not require this information although it is sometimes given. SUPPLEMENTARY TABLES 159 TABLE XXXVII DATA FOR THE CORRELATION OF ILLITERACY AND EARLY AGE OF MARRIAGE, ALL CLASSES3 State Maine........... New Hampshire... Vermont......... Massachusetts.... Rhode Island.... Connecticut...... New York........ New Jersey....... Pennsylvania. Ohio............. Indiana......... Illinois........... Michigan....... Wisconsin....... Minnesota........ Iowa............. Missouri.......... North Dakota..... South Dakota..... Nebraska......... Kansas........... Delaware....... Maryland........ Virginia......... Per Cent* Illiterate, 10 Years of Age and Over 3.3 4.4 3.0 4.7 6.5 6.2 5.1 5.1 4.6 2.8 2.2 3.4 3.0 2.4 1.8 1.1 3.0 2.1 1.7 1.4 1.6 5.9 5.6 11.2 Per Cent of** Females, 15-19 Ever Married State Per Cent* Illiterate, 10 Years of Age and Over - - 11- _ _ - - _ 10.2 Per Cent of** Females, 15-19 Ever Married West Virginia.... 6.4 8..0 North Carolina.. 13.1 9.8 5.0 5.7 6.3 6.5 7.6 10.0 11.5 13.3 9.3 11.4 5.7 5.3 9.4 13.5 6.7 8.0 9.4 12. 1 13.9 12.3 13.7 South Carolina... Georgia......... Florida.......... Kentucky....... Tennessee....... Alabama....... Mississippi..... Arkansas........ Louisiana........ Oklahoma...... Texas........... Montana........ Idaho........... Wyoming........ Colorado........ New Mexico..... Arizona......... Utah............ Nevada......... Washington...... Oregon.......... California...... 18.1 15.3 9.6 8.4 10.3 16.1 17.2 9.4 21.9 3.8 8.3 2.3 1.5 2.1 3.2 15.6 15.3 1.9 5.9 1.7 1.5 3.3 18.4 16.1 17.0 22.9 21.8 20. 1 20.2 20.4 21.1 23.2 18.2 21.1 18.6 10.8 13.1 15.3 13.6 18.9 22.2 10.1 13.8 13.0 12.7 12.0 SMarried, widowed, and divorced. Sources: *Fourteenth Census of the United States, vol. II, chap. XII, Table V; **Ibid., chap. IV, computed from Table XI. 160 ILLITERACY IN THE UNITED STATES TABLE XXXVIII DATA FOR THE CORRELATION OF ILLITERACY AND SIZE OF FAMILY, NATIVE WHITE FEMALES, 15-544 Illiteracy Rate* Per Average Number** of State 1,000 Females Living Children Per Mother of 1920 California.................. Connecticut................. Indiana..................... K ansas...................... K entucky................... Maryland.................. M ichigan.................... M innescta................... Nebraska.................... N ew York................... North Carolina............... O hio........................ Oregon...................... Pennsylvania............... South Carolina............... Utah........................ Vermont.................... Virginia..................... W ashington.................. W isconsin................... 3.6 3.8 6. 7 3.6 52.1 10.8 4.2 3.3 2.8 3.7 66.4 4.9 2.3 5.2 60.0 2.4 8.9 41.6 1.9 4.5 2.2 2.3 3... 2.7 2.8 3.1 2.7 2.6 2.8 3.0 2.3 3.3 2.5 2.4 2.7 3.2 3.1 2.8 3.2 2.4 2.8 4Sources: *Fourteentl Census of the United States, vol. II, chap. XII, computed from Table XII; **Birth Statistics for the Birth Registration Areas of the United States, 1920, p. 15. SUPPLEMENTARY TABLES 161 TABLE XXXIX DATA FOR THE CORRELATION OF ILLITERACY AND URBANIZATION, ALL CLASSES, 10 YEARS OF AGE AND OVER5 Per Cent e ePer Cent Per Cent State Illiter- of State Illiter- of ate* baniza-ate* banization tion Maine............ New Hampshire... Vermont......... Massachusetts.... Rhode Island.... Connecticut..... New York........ New Jersey....... Pennsylvania.... O hio............. Indiana.......... Illinois........... Michigan......... Wisconsin........ Minnesota........ Iow a............. Missouri.......... North Dakota..... South Dakota..... Nebraska......... Kansas........... Delaware......... Maryland........ Virginia.......... 3.3 4.4 3.0 4.7 6.5 6.2 5.1 5.1 4.6 2.8 2.2 3.4 3.0 2.4 1.8 1.1 3.0 2.1 1.7 1.4 1.6 5.9 5.6 11.2 39.6 62.8 31.6 94.7 97.4 67.7 82.6 78.3 65.4 64.6 51.4 68.6 61.9 48.4 45.5 37.9 48.7 14.8 17.1 32.8 36.2 54.6 61.2 31.3 West Virginia.... North Carolina... South Carolina... Georgia......... Florida.......... Kentucky....... Tennessee...... Alabama........ Mississippi..... Arkansas........ Louisiana........ Oklahoma....... Texas........... Montana........ Idaho........... Wyoming........ Colorado........ New Mexico..... Arizona......... Utah............ Nevada......... Washington...... Oregon.......... California....... 6.4 13.1 18.1 15.3 9.6 8.4 10.3 16.1 17.2 9.4 21.9 3.8 8.3 2.3 1.5 2.1 3.2 15.6 15.3 1.9 5.9 1.7 1.5 3.3 27.1 20.8 19.2 27.6 38.6 28.6 28.3 23.6 14.6 18.1 37.5 28.7 34.7 33.0 28.9 30.8 50.9 19.1 36.5 50.2 20.3 56.7 51.4 69.2 5Sources: *Fourteenth Census of the United States, vol. II, chap. XII, Table XII; **Ibid., computed from chap. III, Table XIII and chap. XII, Table XXII BIBLIOGRAPHY I. GENERAL Anderson, J., "The Falling Birth-rate," Nature, Vol. I, (1913), 84-85. Ayers, Leonard P., An Index Number for State School Systems, Russell Sage Foundation, New York City, 1920. Baber, Ray E., and Ross, Edward A., Changes in the Size of American Families in One Generation, (University of Wisconsin Studies in the Social Sciences and History, No. io), Madison, I924. Birth Statistics for the Birth Registration Area of the United States, 1920, U. S. Bureau of the Census, Dept. of Commerce. Birth, Stillbirth, and Infant Mortality Statistics of the Birth Registration Area of the U. S., I924, 1925, U. S. Bureau of the Census, Dept. of Commerce. Bloch, Louis, "Results of Two Years' Operation of the Literacy Test for Admission of Immigrants," Q. Pub. Am. Stat. Assn., XVII, (I920-I92I) pp. 333-35. Boas, F., The Census of the North American Indians, Am Ed. Assn. N. S. I899. Brentano, L., "The Doctrine of Malthus and the Increase of Population During the Last Decades," Economic Journal, XX, (I910), 371-93. Canada, Dominion of, Dominion Bureau of Statistics, Illiteracy and School Attendance in Canada, Ottawa, Canada, I926. Cavan, R. S., Suicide, Chicago, University of Chicago Press, 1928. Chapman, J. C., and Wiggins, D. M., "Relation of Family Size to Intelligence of Offspring and Socio-economic Status of Family," Pedagog. Sem. XXXII, (1925), 414-21. Children's Bureau, U. S. Dept. of Labor, Infant Mortality, 164 BIBLIOGRAPHY A Study of Infant Mortality in a Suburban Community, Montclair, N. J., Series No. 4, Pub. No. i, pp. 1-36, I9I5. Crum, F. S., "The Marriage Rate in Massachusetts," J. Am. Stat. Assn., IV (Dec., I895), 332-39. Dempsey, Mary V., Infant Mortality in Brockton, Mass. Bureau Pub. No. 37, Children's Bureau, U. S. Dept. of Labor. Dublin, L. J., "Infant Mortality in Fall River, Mass.," J. Am. Stat. Assn., XIV (1915) 505-20. Duke, Emma, Infant Mortality, Results of a Field Study in Johnstown, Pa., Based on Births in one Calendar Year. U. S. Dept. of Labor Infant Mortality Series, Bureau Pub. No. 9, I915. Eighth Census of U. S., Vol. IV. Eleventh Census of U. S., Compendium, Part III. Falk, I. S., ThePrinciples of Vital Statistics, Philadelphia and London, W. B. Saunders Company, I923. Farr, W., Vital Statistics, Part III, London, 1885. Fourteenth Census of U. S., Vols. II and VIII. Frenay, A. D., The Suicide Problem in the United States, Boston, R. Badger. I927. Garis, Roy L., Immigration Restriction, New York, The Macmillan Company, I927. Groves, Ernest R. and Ogburn, William F., American Marriage and Family Relationships, New York, Henry Holt & Company, I928. Haley, Theresa S., Infant Mortality. Results of a Field Study in Akron, Ohio, Based on Births in One Year, U. S. Children's Bureau, Infant Mortality Series No. I, Bureau Pub. No. 72, I920. Hart, H., Selective Migration as a Factor in Child Welfare in the U. S. U. of Iowa Studies in Child Welfare, Vol. I, No. 7, pp. 55-56., Differential Fecundity in Iowa, A Study in Partial Correlation. U. of Iowa Studies in Child Welfare, Vol. II, No. 2. Hill, J. A., "Comparative Fecundity of Women of Native and Foreign Parentage in the U. S.," J. Am. Stat. Assn., XIII (1913), 583-604. BIBLIOGRAPHY 165 Hoffman, Frederick L., Suicide Problems, Newark, New Jersey, Prudential Press, I928. Holmes, S. J., "Size of College Families," Jour. Heredity, Oct., I924, pp. 406-I5 -Hunter, Estelle B., Infant Mortality, U. S. Dept. of Labor, Children's Bureau Publication No. 29. Kiser, A. N., "Attempt at a Statistical Determination of the Birth-Rate in the U. S.," J. Am. Stat. Assn., XVI (I919), 442-57~ Leigh, Edwin, Illiteracy in the United States. Annual Report of the Commissioner of Education, I870, pp. 467 -502. Leven, Maurice, Income in the Various States, Its Sources and Distribution, 1919, 1920, and 1921, New York, National Bureau of Economic Research, Inc., I925. Methorst, H. W., "Survey of Birth-Rates of the World," Eng. R., XIX (I927), 116-27. Miner, J. R., "Suicide and Its Relation to Climatic and Other Factors," Am. Jour. of Hyg., Baltimore, 1922, No. 2 of the Monographic Series of the Am. Jour. of Hyg. Morselli, Henry, Suicide, An Essay on Comparative Moral Statistics, New York, D. Appleton & Company, 1882. Mortality Statistics, U. S. Bureau of the Census, Dept. of Commerce, 1919, Table 8; 1920, Table 8; 1921, Table 8. Newsholme A., and Stevenson, T. H. C., "The Decline of Human Fertility in the United Kingdom and Other Countries as Shown by Corrected Birth-Rates," Jour. Roy. Stat. Society, LXIX (I906), 34-87., "An Improved Method of Calculating BirthRates," Jour. of Hyg. V (April and July, 1905). Ninth Census of U. S., Vol. I. Ninth Census of U. S., Population and Social Statistics, XXX. Ogburn and Groves, American Marriage and Family Relationships, Part II. Park, Robert E., "The City: Suggestions for the Investigation of Human Behavior in the City Environment," Am. Jour. of Sociology, XX (1915), 577-612. 166 BIBLIOGRAPHY Parkinson, W. D., "Literacy and the Immigrant," J. Educ. XXX (1914), 567-70. Pearl, Raymond, and Ilsley, Morrill L., "Preliminary Discussion of the Correlation Between Illiteracy and Mortality in American Cities," Am. Jour. of Hyg. II (I922), 587-600. Redford, Arthur, Labor Migration in England, i800-i85o, New York, Longmans, Green & Co., Ltd. Rochester, Anna, Infant Mortality, Results of a Field Study in Baltimore, Md., Based on Births in One Year, U. S. Dept. of Labor, Children's Bureau Pub. No. I19, pp. 1-400, 1923. Ross & Baber, See Baber, Ray E., and Ross, Edward A. Seventh Census of U. S., p. Ixi. Seventh Census of U. S.: Compendium, Part III. Sixth Census of the United States: Illiteracy. Smith, Mary R., "Statistics of College and Non-College Women," Jour. Am. Stat. Assn. VII (900o). pp. 1-26. Sorokin, Pitirim, Contemporary Sociological Theories, New York and London, Harper & Brothers, I928., Social Mobility, New York and London, Harper & Brothers, 1927. Stearns, A. W., "Suicide in Massachusetts," Mental Hyg., V (192I), 752-77. Steele, Glenn, Infant Mortality in Pittsburgh, Series No. 12, U. S. Children's Bureau, Pub. No. 86, pp. 1-24, 192I. Stevenson and Newsholme, "The Decline of Human Fertility in the United Kingdom and Other Countries as Shown by Corrected Birth-Rates," Jour. Stat. Society, LXIX (I906), 34-87. --, "An Improved Method of Calculating BirthRates," Jour. of Hyg. V (April and July, I905). Talbot, W., Adult Illiteracy, U. S. Bureau of Ed. Bul. No. 35, I916. Texas, University of, A Report on Illiteracy in Texas, Bulletin, 1923. Thirteenth Census of U. S., Vol. I. U. S. Bureau of the Census: Illiteracy in Foreign Countries, Dept. of Commerce, pp. I-4. U. S. Bureau of the Census: Marriage and Divorce, I922. BIBLIOGRAPHY 167 U. S. Bureau of the Census: Patients in Hospitalsfor Mental Disease, I923. U. S. Bureau of the Census: Prisoners, 1923. U. S. Bureau of Education: Illiteracy in the U. S. and an Experimentfor its Elimination, Bul. No. 20, I913. Warren, C., Illiteracy in the U. S. in 1870 and I88o, U. S. Bureau of Ed. Cir. of Inform. No. 3, 1884. Whipple, George C., Vital Statistics, An Introduction to the Science of Demography, 2nd ed., New York, John Wiley & Sons, Inc., I923. Wiggins and Chapman, "Relation of Family Size to Intelligence of Offspring and Socio-Economic Status of Family," Pedagog. Sem. XXXII (1925), 414-21. Willcox, W. F., Illiteracy in the United States, i905, Bureau of the Census Bulletin, No. 26. Woodbury, Robert M., "Infant Mortality Studies of the Children's Bureau," Pub. Am. Stat. Assn., XIX (I918), 30-53. Young, A. A., "The Birth-Rate in New Hampshire," Jour. Am. Stat. Assn., XIX (I905), 263-8I. Young, A. A., "The Census Age Question," Pub. Am. Stat. Assn., XII (I910-II), 362. Yule, G. Udny, "On the Changes in the Marriage and Birth-Rates in England and Wales During the Past Half-Century," Jour. Roy. Stat. Society, LXIX (I906), 88-132. II. STATISTICAL Chaddock, Robert Emmet, Principles and Methods of Statistics, New York, Boston, Houghton Mifflin Company, 1925. Crum, William L., and Patton, Alson C., An Introduction to the Methods of Economic Statistics, Chicago and New York, A. W. Shaw Company, 1925. Day, Edmund E., Statistical Analysis, New York, The Macmillan Company, 1925. Dittmer, Clarence G., Introduction to Social Statistics, Chicago and New York, A. W. Shaw Company, 1926. Groves, Ernest R., and Ogburn, William F., American Marriage And Family Relationships, Chapter XXIX, "A 168 BIBLIOGRAPHY Study in Correlation," New York, Henry Holt & Company, 1928. Jerome, Harry, Statistical Method. Chapter XV, New York and London, Harper & Brothers, I924. Kelly, Truman L., Tables to Facilitate the Calculation of Partial Coefficients of Correlation and Regression, U. of Texas Bul. No. 27, 1916. Mills, Frederick Cecil, Statistical Methods, New York, Henry Holt & Company, I924. Miner, John Rice, x/ I -r2 and I - r2for use in Partial Correlation and in Trigonometry, Baltimore, The Johns Hopkins Press, I922. Neifeld, M. R., "A Study of Spurious Correlation," Jour. Am. Stat. Assn., XXII, 33 -38. Pearson, Karl, The Grammar of Science, Part I, 3rd ed., New York, The Macmillan Company, I9II. Ross, Frank A., School Attendance in 1920, U. S. Bureau of the Census Monograph V, I924. Social Science Research Council Publications: Research Method and Procedure in Agricultural Economics, II, 272-73, 1921. Thomas, Dorothy Swaine, "Statistics in Social Research," Am. Jour. of Sociology, XXXV (July, 1929), 1-17. Thurstone, L. L., The Fundamentals of Statistics, New York, The Macmillan Company. I925. Tolley, H. R., and Ezekiel, M. J. B., "A Method of Handling Multiple Correlation Problems," Jour. Am. Stat. Assn., XVIII, 993-I003. Whipple, George C., Fital Statistics, An Introduction to the Science of Demography, 2nd ed., New York, John Wiley & Sons, Inc., I923. Yule, G. Udny, An Introduction to the Theory of Statistics, 8th ed., rev., London, Charles Griffin & Company, 1927. -----, "On the Interpretation of Correlations between Indices or Ratios," Jour Roy. Stat. Society, LXXIII (1910), 644-47. THE UNIVERSITY OF NORTH CAROLINA SOCIAL STUDY SERIES UNDER THE GENERAL EDITORSHIP OF HOWARD W. ODUM. BOOKS MARKED WITH * PUBLISHED IN COOPERATION WITH THE INSTITUTE FOR RESEARCH IN SOCIAL SCIENCE. BECKWITH: Black Roadways: A Study of Folk Life in Jamaica..........3.00 BRANSON: Farm Life Abroad....................................... 2.00 *BREARLEY: Homicide in South Carolina.....................In preparation *BROWN: Public Poor Relief in North Carolina........................ 2.00 *BROWN: State Highway System of North Carolina............In preparation *BROWN: State Movement in Railroad Development..................... 5.00 CARTER: Social Theories of L. T. Hobhouse........................... 1.50 CROOK: General Strike, The........................................ 6.50 FLEMING: Freedmen's Savings Bank, The............................. 2.00 GEE (ED.): Country Life of the Nation, The.......................... 2.50 *GREEN: Constitutional Development in the South Atlantic States, 1776 -1860.......................................................... 4.00 GREEN: Negro in Contemporary American Literature, The.............. 1.00 *GRISSOM: Negro Sings a New Heaven, The........................... 3.00 HAR: Social Laws............................................... 4.00 *HEER: Income and Wages in the South.............................. 1.00 *HERRING: History of the Textile Industry in the South..........In preparation *HERRING: Welfare Work in Mill Villages............................ 5.00 HOBBS: North Carolina: Economic and Social......................... 3.50 *JOHNSON: Folk Culture on Saint Helena Island....................... 3.00 *JOHNSON: John Henry: Tracking Down a Negro Legend............... 2.00 *JOHNSON: Social History of the Sea Islands.......................... 3.00 JORDAN: Children's Interest in Reading............................... 1.50 KNIGHT: Among the Danes......................................... 2.50 Lou: Juvenile Courts in the United States............................ 3.00 *METFESSEL: Phonophotography in Folk Music....................... 3.00 M ILLER: Town and Country........................................ 2.00 *MITCHELL: William Gregg: Factory Master of the Old South............ 3.00 *M URCHISON: King Cotton is Sick................................... 2.00 NORTH: Social Differentiation....................................... 2.50 ODUM: Approach to Public Welfare and Social Work, An............... 1.50 *ODUM (Ed.): Southern Pioneers.................................... 2.00 *ODUM and WILLARD: Systems of Public Welfare...................... 2.00 *ODUM and JOHNSON: Negro and His Songs, The..................... 3.00 *ODUM and JOHNSON: Negro Workaday Songs....................... 3.00 POUND: Law and M orals........................................... 2.00 *PUCKETT: Folk Beliefs of the Southern Negro......................... 5.00 *RHYNE: Some Southern Cotton Mill Workers and Their Villages....... 2.50 ROSS: Roads to Social Peace...................................... 1.50 SALE: Tree Named John, The...................................... 2.00 SCHWENNING (Ed.): Management of Problems......................... 2.00 SHERRILL: Criminal Procedure in North Carolina...................... 3.00 *STEINER and BROWN: North Carolina Chain Gang, The.............. 2.00 *VANCE: Human Factors in Cotton Culture.......................... 3.00 *WAGER: County Government and Administration in North Carolina..... 5.00 WALKER: Social Work and the Training of Social Workers.............. 2.00 W HITE: Some Cycles of Cathay...................................... 1.50 WILLEY: Country Newspaper, The................................... 1.50 WINSTON: Illiteracy in the United States............................ 3.50 The University of North Carolina Press, Chapel Hill, N. C.; The Baker and Taylor Co., New York; Oxford University Press, London; The Maruzen Company, Tokyo; Edward Evans & Sons, Ltd., Shanghai. MUTILATE CARD (JA F, *"II.-I --- I