Autlior . ^t*o^ o z '^*v."i^ Title 8.51 Imprint . U — »71IT»-» %ra UNIVERSITY OF PENNSYLVANIA EDUCATION AND FECUNDITY BY NELLIE SEEDS NEARING ABSTRACT OF THESIS PRESEN rED TO THE FACULTY OF THE GRADUATP. SCHOOL IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY REPRINT FROM PUBLICATION OF AMERICAN STATISTICAL ASSOCIATION JUNE, 1914 CHAUTAUQUA PRINT SHOP CHAUTAUQUA, N. Y. 1917 UNIVERSITY OF PENNSYLVANIA EDUCATION AND FECUNDITY BY NELLIE SEEDS NEARING ABSTRACT OF THESIS PRESENTED TO THE FA<;ULTY OF THE GRADUATK SCHOOL IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY REPRINT FROM PUBLICATION OF AMERICAN STATISTICAL ASSOCIATION JUNE, 1914 CHAUTAUQUA PRINT SHOP CHAUTAUQUA, N. T. 1917 W'> ?^^ a\ Gift EDUCATION AND FECUNDITY. I'.Y NELLIE SEEDS NEARING. 1. The Increase in the Higher Education of Women. In the effort to analyze a popular prejudice or fallacy, an ade- quate investigation of facts is usually sufficient to reveal the origin of such prejudice in previously existing conditions which have altered so gradually with time, that society as a whole has hardly yet become aware of the change. In such a category we might include many of the fallacies concerning the position and faculties of women. A hundred years ago. the higher education of women wa< an unheard of phenomenon. Women were intended to bear children. What need had Ihey of an education! Marriages were then con- tracted at an early age, usually in the teens or early twenties, and as an advanced education precluded the possibility of an earlv marriage, it seemed to preclude possibility of marriage at all. The average woman, therefore, who went to college in the early days of college training for women, was not the type who would have been apt to marry in any case. The first classes of college women consisted largely of the woman who had some spe- cial talent which she wished to develop and practice, the woman of strong intellectual proclivities, who preferred not to engage in the domestic occupations usually relegated to women, and the woman who, because of personal unattractiveness, knew or feared her lack of popularity among men. The later experiences of women's colleges are quite the opposite of those early encountered. Today it is the normal, not the un- usual girl who goes to college or technical school. College edu- cation is considered generally desirable as a means of finishing the education of the average woman. It gives culture. It has be- come a common comfort, if not a necessity, for young women of means. The developments of the last twenty years have placed before the parents of all classes a powerful incentive to give their daughters the best in education that can be secured. Are parents today recognizing these facts and educating their daughters as well as their sons? What are the facts? .'-Vfter thev have been EDUCATION AND FECUNDITY. ascertained, the second question may be asked — Does this increaNC of education have any appreciable effect upon fecundity? The facts point conckisively to a rapid advance in the higher education of women. In the first place a larger number of girls than of boys are being given a high school education in the cities of the United States. Among 318 cities of the United States, 132 having 25,000 population and over. 186 having less than 25,000 population, twenty report more boys than girls in the fourth year of the high school. These twenty, moreover, are small and com- paratively unimportant. In all the other 298 cities the number of girls is uniformly greater than that of boys.* Although the number of girls in the high schools almost uni- versally exceeds the number of boys, the situation in the elemen- tary grades is the reverse. A school census of thirty-two state-^ (1911) shows that in only one state — New Hampshire — were there more girls than boys in all grades of the sclionlst. In orJv eiglit nut of these thirty-one states, however, did the difference between the number of boys and number of girls exceed 10,000. The school system in its entirety contains a higher proportion of boys than of girls. It is only in the high school grades that the number of girls exceeds the number of boys. Clearly then, girls are availing themselves more fully of the public higher educa- tional opportunities than are boys. ■ The real extent of the entrance by women into the fields of higher education is shown by a study of the statistics of college students. The movement toward the college education of women is so recent, and the opportunities for such education so much smaller than for that of men, that some years must still elapse before the absolute number of girls in college aijproaches the absolute number of boys there. The vital question therefore is the rale of increase in the number of women attending college each year and the rate of increase in the number of men. Tlic figures are available between 1889- 191 2 for the number of men and women in colleges of the United States.* .A comparison of * Age and grade census of Schools and Colleges, G. D. Straker, United .States Bureau of Education Bulletin 191 1, No. 5, pp. 14-28. t Report of United States Commissioner of Education, 1911-1912, Gov- einmcnt Printing Office, igi,'? Vol. 2, p. 93. EDUCATION AND FECUNDITY. these figures, first by ten year periods, then for the total period, shows that the ratio of increase is far greater for the women than for men TABLE I. PEK CENT. OF INCREASE IN THE NUMBER OF MEN AND WOMEN IN THE COLLEGES OF THE UNITED STATES. 1889- 1912. Period. Men. Women. iSfjo-iSgi ot 1900-1901. 1900-1901 to 1910- igi I. 1889-1890 to 1911-1912. 64 77 S8 66 179 . 248 A further examination of the figures shows that the proportion of men to the total number of college students is steadily decreas- ing, and inverseh- the proportion of women steadily increasing. T-\BLE II. PROPORTION OF MEN IN COLLEGE TO TOTAL COLLEGE STUDENTS. 1889-1912. Period. Per Cent. 1*301891 1.J00-1901 lyio-igt 1 1911-1912 68 66 64 63 Here then is an answer to the first question regarding the higher education of women. The number of women in the United .'states receiving higher education is absolutely and relatively in- creasing in the colleges, while in the high schools the number of girl students actually e.xceeds the number of boy students. What effect, if any, has this general entrance of women into the fields of higher education upon the marriage and birth rates of the women? An answer, scientificalh- established, will put to rest popular i^rejudice and s])eculative theor-\'. *Total number of men and women in the colleges uf the United States, 1889- 1912. Report of the Commissioner of Education, 1911-1912, Govern- ment Printing Office, 1913, Vol. 2, p. 249. 1889-1890 44.9-26 20,874 1890- 1891 46,220 22,03b 1900-1901 '. 75.4/- .38,900 1910-191 1 119,026 64,546 1911-1912 125.750 72.703 EDUCATION AND FECUNDITV. II. Marriage Rates of Educated Women. Marriage is the first element in fecundity which must be con- sidered and investigated. Illegitimate births are few among the educated classes, and the statistics of such births even in the population at large, never exact, probably do not even approxi- mate the truth. Hence this discussion will ignore the whole question of illegitimacy, and consider marriage as the necessarv precursor to motherhood. The method adopted in obtaining the statistics of marriage among college women was as follows: From a list of all the colleges in the L'nited States, classified according to states, in the report of the United States Commissioner of Education for 1911-1912, showing in all 459 colleges liaving wtnnen students, seventy-two colleges were selected from twenty-eight states, an effort being made ttj select those having the largest number of women students. The women in these colleges are representative, b(jth territorially and numerically, df the college women of the United States. The colleges were selected from various localities. The tot;d number of women students in these seventv-two colleges was 40,653 or 36 per cent, of the total number of women college students in the United States. .•\ form letter was written to the registrar of eacli of these colleges asking for any vital statistics which had been collected, or which might be available in any form. Twenty-nine colleges rei-ilied that n(5 such statistics were available, eighteen sent Uii reply, and twenty-five colleges sent such information as they had, which information in seven cases proved of no value. The ma- terial at hand was then carefully gone over and compiled into marriage rate and birth rate tables for each college. The marriage tables for each college were then combined into two tables giving by decades first the number and second the per cent, of married and unmarried graduates from eighteen colleges lun'ing a total of 14,551 women graduates. EDUCATION AND FECUNDITY. 5 TABLE III. •NUMBERS OF WOMEN GRADUATES AND NUMBER OF SUCH GRADUATES MARRIED FOR CERTAIN COLLEGES. BY COLLEGE AND BY DECADES, 1870-1913. 187c •79 1880-89 1890-99 1900-09 1 1910-12 Total to 1 1012 College. g-i w 0- 0*1 02; D.3 c 0- Si oz C.3 n gS! ^3 OZ it CO* ft gz G.-1 c o- 0.1 EarUiam Swarthmore VV'iihion Indiana Vassar Oregon Radcliffe ... Wellesley.. S. Dakota... iJryn Mawr Mississippi. Holyoke .... Washington State Rochester ... Ohio Oregon Agri- cultural ... kockford . . . Sini'ih 23 t 13 21 20 30 203 43 1% 56 376 34 3l8 ■ 2' 33 42 23 40 208 19 12 2S7 7 '3? 148 219 228 791 84 253 1201 47 294 10 377 80 86 126 120 52.1 37 123 3 158 zoo 260 3f-9 482 1874 '51 746 733 688 48 82 .35= 103 103 164 500 63 228 33 244 8 34 30 ■5 J09 r4u 228 23 83 10 4 2 16 5 13 746 887 3424 269 1257 1719 153 1 193 70 174 135 52 398 237" 2l6l» 266 366 1287 136 376 780 84 392 18 56 40 32 206 134 1016 TABLE IV. RER CENT. OF WOMEN COLLEGE GRADUATES WHO MARRIED, FOR CERTAIN COLLEGES, BY COLLEGES AND BY DECADES. 1870-1913. College. Earlham Swarthmore Wilson , Indiana Vassar Oregon Radcliffe \\'eHesley South Dakota Bryn Mawr Missis'iiiipi Holyoke Washington State Rocliester Ohio Oregon Agricultural. Rockford Smith 1870-79 56.9 67.7 5.S.5 85.7 53.0 I l8u- 89 76.7 58.3 54-7 71-4 55-3 55-8 40.0 49.6 Per Cent, of Graduates Married 1890-99 Total Gradu- ates to 1912 60.6 58.7 56.1 55-2 47.5 64.2 47-4 43.5 78.7 41.8 30.0 41.9'' 51.5 39-6 30-7 ,!4.9 26.6 41.7 30.5 21.7 35-4 16.6 23.6» 41.4 22.2 I ■3-3 3.6 7.6 T5.6 50.3 42.2 35.6 41.2 37-5 50.5 29.9 45-3 54-9 32.8 25.7 32.1 22.2 61.S 51.7 56.1 47.0 2 Since 1873. ^Since 1883. = Since 1901-10. *Since 1873. ° Since 1885. « To 1901. _t Secured from an article by Amy Hewes, Quarterly Publication of American Statis- tical Association, Vol. XII. p. 771. 6 EDUCATION AND FECUNDITY. In only live cases were figures available for the decade 1870- 79. ill only nine cases were they available from 1880-89. In view of the fact, further, that the marriage rate of those graduat- ing since 1900 would still necessarily be low because such gradu- ates would be under 35 years of age, the most representative decade under consideration would be that from 1800-99. Here the absolute number of women graduates is greater than in any precefling decade, although less than in any succeeding decade. The per cent, of marriage here averages, on the whole, consider- ably lower than that of the previous decades in tlie colleges from which such information is available. It is probable, therefore, that the per cent, in this decade would give a conservative estimate of the rate for college women. It varies from 78.7 in South Dakota State College to 30 per cent, in the L'niversity of Mississ- ippi^neither of these colleges being representative, graduating between them in the decade only 57 students. If we compare the per cent, for this decade with that of the following, we find an average fall of 20 while the average fall from the per cent, of 1900-09 to that which was estimated in a few cases for 1909-1912 was 2^ per cent. The totals are comparatively meaningless. This statement applies to both the ab.solute number and to the )M"oportion of graduates married. Pioth the total number of graduates married in each college ( and in some colleges this was the only information available ), and the proportion married must be considered in view of the fact that the results are obtained from a body of students of Whom a large number have yet had little o]>portunity for mar- riage, and whose marriage rate when estimated separately wis from 20 to 23 per cent, lower than that in the previous decade. It must be remembered, moreover, that as each graduating cl:is> is in almost every case larger than the preceding one. the denom- inator of the fraction is every year being increased at a greater rate than the numerator. On the other hand, this marriage per cent, of the entire number of graduates of a college includes in tlie older colleges those who have graduated in the decades previous to 1890, and whose marriage rate averages, where figures are ob- tainable, over 15 per cent, higher in 1880-89 than in 1890-99 and EDUCATIOy AND FECUNDITY. J 5 per cent, higher in 1870-79 than in 1880-89. Balancing off tlie 20 and 25 per cent, fall in the rate from 1890 to 1913, and 15 and 5 per cent, increase previous to 1890 and taking into account that the fall was for a far greater absolute number than the increase, because of the yearly increase in size of classes, we may assume that in cases where the marriage rate by classes was not obtain- able, the rate for those graduating in the decade 1890-99 is at least 10 per cent, higher than the rate for the entire body of graduates. The earliest years, as already indicated, are scarcely repre- sentative. Yet there seems to be little real dift'erence between the percentages there and in later decades. Of the 9 colleges which supply us figures for the decade 1880-89, 2 show percentages of married graduates in excess of 70 per cent. In neither of these colleges was the total number of graduates large ( Earlham 43, Indiana 56). The other 7 colleges report marriage rates varying less than 10 points. The 2 large colleges, Vassar and Wellesley (376 and 518 graduates respectively), report a marriage rate for tl'.e decade of 55.3 in the case of \'assar and 49.6 in the case of Wellesley. The decade 1890 to 1899 is undoubtedly the most fairly representative of any of the decades under consideration. On the one hand, it falls within the epoch which accepted College Education for women, and looked upon it as thoroughly respect- able. On the other hand, the graduates in the latest graduating class (class of 1899) are now at least 35 years of age. The mar- riage record of the decade is therefore fairly complete. The 8 colleges graduating more than 100 students during the decade ( Earlham, Swarthmore, Wilson, Indiana. N'assar. Radeliffe. Wellesley, and Bryn Mawr) show fairly uniform marriage rates, the lowest is Bryn Mawr, 41.8 per cent. (294 graduates), and the highest is Swarthmore, 58.7 per cent. (148 graduates). It is probable that the marriage rate for this decade is fairly represen- tative of the tendency in the modern women's college world. Granted that this statement is accurate, it ma}- be said that the proportion of women college graduates who marry is a]i]5roxi- mately one half (u.sually slightly over 50 per cent.). The records of the succeeding decade show a heavy falling off in i)er cent, of women marrieii. The women in the class of 1909 EDUCATION AND FKCUNDITY. are approximately 25 years old. The marriage records are there- fore incomplete and wholly unreliable. The proportion married for this decade i.s from one quarter to one third of the total graduates, and is therefore almost one half of the pupils reported for the previous decade. This study was originally intended to stu(h' in three classes: ( 1 ) The college woman, ( 2 ) The normal school graduate, and ( _^ I The commercial high school graduate, and to compare the mar- riage and birth rates prevailing among the three. With this end in view a form letter similar to that sent to the colleges was sent to leading high and normal schools throughout the country. (')iily six replies were received and in only 2 of these cases was any in- formation vouchsafed. Obviously these figures were of no pr;'.c- tical use. The only other group educationally (in a par with women col- lege graduates, for whom statistics are obtainable, is the grouj) of men college graduates. The ]'vle Rcticzv gives us the following marriage rate for Yale graduates: TAin.F. \'. PER (EXT (IF V.1.I.E (;R.\I)1 AXES WHO WERE MAKKlEl). irui l.S8^. Classes. i Per Cent. Married I 1701-1701 S8.3 1707-1833 *•* 1834-1849 ' R1.2 1849-1866 1 8i..i 1867-1886 1 66.3 The figures for the classes 1867-1886 are the only ones in any way com|)arable with the figures for woman college gradiuates. Since the question of social standing did not enter into the m:ir- riage rate of college men during the earlier decades, their mar- riage rate for the earliest decades would be expected to be some- what higher than that of college women. The only remaining comparison to be made is that lietween the marriage rate of college women and the marriage rate nf the ])opulation at large. EDUCATION AND FECUNDITY. TABLE VI. PER CENT. MARRIED.- TOTAL NUMBERS AND PERCENTAGES OF WOMEN. IS YEARS OF AGE AND OVER, TOGETHER WITH THOSE WHO ARE MARRIED AND WHO ARE SINGLE. 1910. Total. Married Number. ^^^ Cent. Number. Per Cent. Total: Native white 30,047, ,525 I 100 Native parentage 15,523,900 1 100 Native white, foreign or I I mixed parentage I 5,887,131 too Foreign-born white 1 5.446,306 1 100 Negro I 3.103,344 I mo 21.045,983 10.842,998 3,421,147 4,444,657 2,269,066 J_ 70.0 6y.8 58.1 81.6 73-1 Single Number. 8.933.170 4,644,122 2,453.017 994. 1 1 823,996 Per Cent. 29-7 29.9 41-7 18.3 26.6 .Since the great majority of women college graduates are native- born, the only figures here with which comparison can be made are those referring to native whites of native parents. As au- thority for this assumption, I miglit (|uote an investigation cover- ing 1.290 college graduates in wiiich 705 replied, practically all of wlioni were native born : 83.3 per cent, had native Aimerican parents. 2.8 per cent, had foreign mother and native father. 3.5 per cent, had foreign father and native mother, only 9.8 per cent, had both parents fcjreign born. The following figures then stand ccmtrasted : ll'aiiioi 75 Years and Over. Per Cent. Married. (native white, native parents )... .69.8 (figures taken 1910). College graduates to 1913 42.2 ( figures taken 1913 ). In view of the fact that a small proportion of college graduates (about 15 per cent. ) are of foreign or maxed percentage, and that the marriage rate for tlie population at large for native white women of foreign parentage is 58.1 ])er cent., hence lower than that for native white of native parentage, the 42.2 ])er cent of college graduates should really be compared with a figure ob- tained b\- including 15 per cent, of those with foreign or mixed parentage, tluis lowering the rate slightly. In other words, if we add to the 15,525,900 of native white jiarentage 15 per cent, of the 5.887,131 of mixed parentage, or 883.069 and to the 10,842,998 ♦Abstract of the Thirteenth Census. 10 KDL'CATIOX AND FECUNDITY. married of native white parentage. 58 ]3er cent, (the foreign parentage rate) of this 883,069, or 158,925. we get a total of 11,001,950 women married out of total 16,406.969. This would give a marriage per cent, of 67. We can then make our contrast between 67 per cent, for the whole population and 42.2 per cent, for college graduates. These figures show that the marriage rate for non-college women is 59 per cent, higher than the rate for college women. This would seem on the face of it, to be a tremendous difference, but it cannot be attributed solely to the difiference in etlucation. There are two other factors that play a large part in the situation. First comes the financial situation. In the second place. soci;d considerations are of the utmost importance in determining mar- riage and usually act in actual practice as deterrents rather than incentives. The standards of living among educated people are higher, and their requirements harder for the prospective husbaiMJ to meet. These two factors might and do cimtrlbute tcward the comparatively low marriage rate of college women. No other available figures throw light on the marriage rate, actual or comparative, of college wnmen. College women do marry, probably in fifty cases out of ;i hundred, given sufficient time out of college. 111. KlRTH RaTE.S OK EoiKArED VVOMIi.N. Like the statistics of marriage, the most satisfactory statistics of the birth rate among educated women are those nf college graduates. Out of the 58 replies from the 71 colleges written to, only 5 vouchsafed any information from which ;i liirth rate could be deduced. In addition, information concerning 4 other college? was obtained indirectl}'. Thus the source f.-om which the figu^c.^ were drawn is small in comparison with the entire field. The colleges included in it are, however, the most representative of American colleges for women. The following table gives a mim- ber of the most pertinent facts about the birth rates of college women : EDUCATION AND FECUNDITY. 1 o a. « < ? ? "> '^ u, •< CD < re Si 3 o o * :t~ o H! ^ us 3 "■' *< 3 n ?r o > -t fi n V =r^ O U rt 3- 5? S "- : ? TT S 1 (T 5! U) ^^ ^ . w V ^ w n" n • ^ i n 1 ; ' 1 •1 o 3 - ^" — ^^^ — •— "l -"^ : ^ ? t «" «■ s ^S' «■ s-^ T o O ■SI ui o ■^ H" < 1 1 i ■ ■ 1 f II 3 3 a. fT • "■ 13 ft ? 5' n 5" 3_ 1 1 ■VI V4 ■VJ i IK ■vl •1 2 !f c a* S 1 , J> 00 00 " o c> a* 1 fi) 3. *^ a a rt u « A ' 4^ ^ Ln 1 a. (t ■^I ^ lj\ ■vj p\ n O b o >J b b Wi 3 c p a :2 K n >d c p H j ^ z c B s a n _ c H 1 5 s ^ IS Cm 3 ■1 3'< ft lis 3ZP OS' 0. s ^ « -p ■fi ^ 1- n n to' [ ^ -^ o « - 3 o' I 3 O 3 1 3 i s 5 ?S: 3 in 3- o ft ' P u =3 ^ table it appears that the per cent, of married college graduates having one or more children varies from 67 per cent, at Smith to 72.3 at Rockford College. In all cases the period covered is approximately the same ( Rryn Mawr is the sole exception ). In all cases the percentage by one or more children is remarkably uniform. The per cent, having two or more children varies from 44.7 at Smith to 53 at Bryn Mawr. The per cent, of married graduates having no children varies from 20.1 per cent, at Rock- ford to 32.9 per cent, at Smith. Thus we find the birth rate is lowest at Smith, highest at Rockford, with Bryn Mawr as a medium. This table of marriage rates and birth rates gives the figures from approximately the beginning of each college down to 1901. In the case of Bryn Mawr and of Holyoke. 'however, the found- ing of them was at a much later date than that of the others. (Holyoke existed as a Seminary only, prior to i8yo. ) Since therefore neither of these colleges has the body of older graduates thnt the other colleges have, we should expct to find marriage and birth rates corresp(indingly lower. Since this is not the case in the birth rates at Br\'n Mawr, we may assume a somewhat higher birth rate there in proportion to the other colleges, than is shown b\' the figures. Since the figures for the 8 colleges whose com- bined records appear show a slightly lower birth rate than the median taken from the entire table, we may assume a slightly lower rate for the colleges included in it. Wells, RadclitTe, North- western, whose se]iarate figures were not obtainable, than for the colleges fnr which separate figures were secured. A comparison of the absolute number of children in each case with the nnml)er of graduates in the 4 colleges from which these figures were obtainable, shows the following results : Smith 59.4, Yassar 83.9, Bryn Mawr 82.3, and Holyoke 76.3 children per 100 graduates. This rate apparently falls far short of properly main- taining the population. The ,s.mall percentage of marriages among graduates, coupled with the low Ijirtli rate spells population 4 7 8 II 11 3 5 6 6 2 3 2 2 : 289 503 135 383 374 It will here be seen in the case of Vassar that the size of fam- ilies has been smaller since 1901 tlian previous to that date. No Vassar graduate since that date has had a family of over three children, and only three have had a family of three. Vassar graduates since 1890 have had no families of six, six families of five, and eighteen families of four children each. The figures of Bryn Mawr, therefore, while they seem lower on the table, are really not so for the same period of years. The last table which seems of value is published in the Btmi Mawr catalogue for 1913, page 298. EDUCATION' AND FECUXDITV. 15 TABLE X. XUMBER OE MARRIAGES AMONG BRYN MAWR GRADUATES "AND NUMBER OF CHILDREN BORN, BY YEAR OF MARRIAGES, 1890-1912. Year of Marriage 1890 1891 1892 ■ 8.JJ 1894 iSyi 1896 1897 iSgS >8W 1900 lyoi 1902 1903 1904 iSf^a lyo(i 1007 I9t.8 1909 191.. 1911 • 912 1890 to 1912. Duration of Mar. Number of Gradu- ates Mar- ried Each Year Number of Children Average per riage in Years Boys Girls Total Marriage 22 .z.i 1 1 2 3 3.0 21-22 4 5 5 10 2.5 2U-2I 2 3 3 1,5 lu-2n ,i 6 8 1 14 4-7 l8-ly 5 J 4 1 9 1.8 ,7-18 9 7 9 1 16 1.8 16-17 3 6 1 J 1 8 2-7 15-16 4 7 3 1 10 2.5 14-15 3 12 9 1 21 4.2 lJ-14 14 10 1 16 I.l 12-13 12 14 1 14 28 2-3 11-12 ij 18 14 ; 32 2.5 10-11 10 13 8 1 21 2.1 9-10 15 11 15 1 26 1-7 3-9 2i 31 32 1 63 2.7 7-8 29 29 ) 24 1 53 1.8 0-7 s6 23 34 1 57 1.6 5-0 29 18 ■ 7 1 35 1.2 4-5 30 22 11 1 33 I.I .i--l 28 16 18 34 i.i 2- J 31 3 15 18 U.6 34 ■ 8 6 14 0.4 I niier 1 S2 u 1 1 u 0.0 392 261 263 524 1-3 We here have an opportunity for the first time to cuinpare the number of diikh-en with the duration of the marriage. The aver- ;\i;e inuiibfr of children ])er family of ten or more years' duration i- 2.7. showing a rate slightl)- higlier than that necessary to maintain a static population, provided no deaths occur, Holyoke gives a similar figure here, showing an average of 2.43 children born to each of 43Q married graduates of the decade 1890-99. Available figures dealing with the fecundity of college women are few in number and narrow in ^cope — the same data are not given in everv case, and coiiiparison is difficult. Figures for the i^ast deal with the number nf graduates having children, rather tli-m the number of children, hence the difficulty of showing ac- curate details of the number of children per marriage. Since in each case, of the five colleges furnishing the best data. Bryn Mawr \v.>; nearest to the medium, her figures can reasonably be as- sumed to represent the approximate situation in those colleges not giving ct)mplete data (.Smith and Rockford). Since further, the l6 ICDUCATION AND FEfUNDITY. Bryn Mawr table for duration of marriage shows that all mar- riages of over ten years' duration average 2.7 children, the fol- lowing conclusion may be drawn : 1. Where all college graduates are included up to date the number of children per graduate would be slightly above three fourths, the number per married graduate would be approximately one and one half. 2. Where only those graduates who have been graduated a sufficient number of jears to allow for marriage antl all probable family increase, are considered, the rate would be api)roximately 2 to 2V^ children per family. So much for the available figures for college graduates. Of all possible comparisons, the fairest, in fact the only fair com- parison, is with the sisters, cousins, and friends of these college women, who did not themselves attend college. Here social and economic considerations would have equal weight, and any differ- ence in rate would most ])robably be traceable to the college training. Mary Roberts Smith in an article on the Statistics of College and Non-College Women (1900)" has made such a study and obtained from it valuable conclusions. Schedules were sent by her to 343 college mothers and 313 non-college mothers who were tlieir sisters, cousins, and friends. She summarizes her con- clusions as follows : 1. The marriage of college women was postp;ined two year< as compared wit'h that of non-college women (26.3 vs. 24.3 years. ) 2. The age of marriage for both classes has been growing in the last thirty years, — a larger per cent, of non-c(illege mothers marrying before the age of 21 ; a larger per cent, of the college mothers marrying after 34. 3. The non-college women have been married an average of 2 ve;irs longer than the college women and have borne a slightl)' larger number of children, but the college women have borne the larger number of children per year of married life. .Signiilcant comparison may be made between the men. If we *Quarterly Publications of the American Statistical .Association, No. 49-50, Volume VII, March-June, IQOO, p. i. EDUCATION AND TI-XU NDITY. compare the fecuiulity of Yale graduates to the fecundity of women college graduates, we get the following results.* TABLE XI. A COMI'AKISON OF THE BIRTH RATES OF VALE AND VASSAR C.RADU- ATE3 CLASSES OF 1867 TO 1886 INCLUSIVE. NUMBER OF CHILDREN PER MARRIED GRADUATE. Vale: 1867-1886 V.iS,-,ar: 2.02 '■9i .\nother interesting comparison might be made between women college graduates and women graduates of normal and high schools, who did not go to college, but here again it is rendered impossible by the fragmentary character of the material. Author- itative data was secured for only two classes and conclusions from them would carry little weight. A comparison of college birth rates with those of the popula- tion at large are of little value because of the diiiferences in age, sex, social and economic position. Moreover, statistics of the population at large in the United States have never been compiled. In the registration area, where some work has been done, imper- fect registration of birth renders the figures of somewhat doubt- ful value. One study, however, has been made vvliich gives us comparable figures from the population at large. I refer to an article by Dr. Jo-eph A. Hill in the Quarterly Publkwtions of the Ameut- CAN Statistical Association.! Dr. Hill's study included in all 185,788 women, 78,432 of whom had been married ten to twenty years. These 78,432 were divided into four classifications : ( i ) white native parentage, (2) white foreign parentage (first gen- eration ) , ( 3 ) ( second generation ) , ( 4 ) negro. Tlie areas covered were the state of Rhode Island, the City of Cleveland and 48 mainly rural counties in Ohio, Atinneapolis, and 2T mainly rural coMinies in iMinnesota. The native white of native parents are of ♦Statistics of Yale Graduates, Yale Review, 1908-09, p. ay. tQuartcrly Pulilications of the American Statistical Association, Vol. XIII. IX-c. 1913, p. 583- F.nUCATIOX AND FECUNDITY. course the only ones that can fairly be com])ared with our college women. For them the fis'ures were as follows ; TABLE XII. BIRTH RATE OF NATIVE-BORN WHITE AMERICAN WOMEN UNDER 4S YEARS OF AGE, MARRIED 10-20 YEARS. Total Bearing- No Children Total Number of Children 42,933 Average [ Per Cent Bearing per Number Number I Per Cent. 1 1 2.(107 n-i 1 Married Woman , q^ ^ 3. 4 or 5 6 or More Children i^.o^.l 2.7 39-8 ■ 37.2 Q-M The interesting fact here is that Dr. Hill's average number of children per married woman, 2."/ , is identical with the 2."] children which each Bryn Mawr marriage of over ten years duration aver- ages. Here, at least, we find two groups of women, college, and selected from the population at large, identical as to age, race, and duratiim of marriage, who average the same number of chil- dren per marriage. In conclusion we may say, hrst, that the birth rate of college women i> ascentainable ; second, that, in so far as it can be coni- jiared to that of the sisters, cousins, and friends uf college women, it is probably verv little lower, or about the same, x.c, the non- college woman has more children, but the college woman bears more per year of married life: third, if we compare it with that of men college graduates, we tind oid\ a slightly lower rate ft.r tlie women ; fourtli. in comparing it with the only definite study so far discovered, of women fmm the pnpulation at large, of similar race, of child-bearing age, married ten tii twenty years, we find almost completelv identical figures. Obviously none of these coni- ])arisons -ire of great value. The numbers included are in mo-t cases tijo small or the ground covered too limited Id enable u.-- to give tiie re>ults any widespread applicatimi. The purpose of this study, as stated in the introduction, was the answering of two definite questions: First, is the- hi.gher education of women absolutely and relatively increasing in this countrv? .Second, has tliis increase, if there is any. an appreciable efi'ect upon fecundity? The first of the-^e (luc-tiims must be answered definitely in the EDUCATION AND FECUNDITY I9 affirmative. A careful investigation of the figures at hand shows not only a larger absolute number of girls than of boys in the high schools of the United States, but a greater ratio of increase in college attendance for girls than for boys, and a steady increase in the proportion of women students to the entire body of college students. The answer to the second question — Has this increase in the higher education of women any effect upon fecundity — is not so satisfactory. It has been found ( i ) that the proportion of women college graduates who marry is approximately one half (slightly over one half in most cases); (2) that the proportion of men college graduates who marry is somewhat higher; (3) that, com- paring college women with the census figures of population at large, the marriage rate among non-college women is consider- ably higher (59 per cent.) than among college women; (4) that the higher marriage rate of men and non-college women (where it is higher) may quite conceivably be due to causes other than that of education. The figures and conclusions obtained are, at best, inadequate. Obviously they fall far short of showing any appreciable effect uf the higher education of women upon fecundity. The college statistics are accurate and reliable as far as they go, but there are too few colleges which keep any records of fecundity, and the greater part of the figures obtainable are too recent to be fairly representative of college fecundity. High and normal school statistics on the subject of fecundity are practically non-existent, ;:nd present a field of work which would adequately repay careful investigation. The figures do prove conclusively the impossibili- ty of justifying any statement that the higher education of women (liies or does not lower fecundity. .\s far as any positive conclusion is concerned, the figures only • show the existence of two separate phenomena of continually increasing importance, which might, were all the facts of the case revealed, show a close inter-relation. Obviously the importance of investigating the facts and obtaining all possible information and statistics cannot be overestimated. The entire trei;d of our higher education of wortien should be. and I confidently believe will be determined in large measure by its eft'ect upon fecundity.