Hunt: College Birth Rate 



57 



(5) It is probable that those who 

 filled out the blanks are more inter- 

 ested, in general, in children than those 

 who did not. A person whose inter- 

 est in children has never been very 

 great might frequently disregard such 

 a questionnaire as unimportant. Such 

 people are not as likely to have off- 

 spring as those who are interested in 

 children. Therefore, there is reason 

 for believing that the number of chil- 

 dren per capita reported is larger 

 than it would have been had the whole 

 group of living graduates returned the 

 questionnaires. 



(G) A determination of the number 

 of children per graduate would have 

 to be based on a group selected in such 

 a way as to have nearly the same per- 

 centages of married and unmarried 

 persons as in the alumni body as a 

 whole, if the determination were to 

 present accurately the facts for the 

 latter. The Alumni Register for 1921 

 shows that 56.9 per cent of all the 

 women graduates in the period 1870- 

 1899 were married. The percentage 

 of married women (excluding those 

 whose complete family size to date was 

 not given) in the data derived from 

 the questionnaires was 56.2 per cent. 

 When all the women reported in the 

 questionnaires (52 married and 32 

 single) are considered, the percentage 

 of married is 61.9 per cent. This 

 shows that the ratio of married to un- 

 married women in our sample is not 

 far from the corresponding ratio for 

 the whole group of graduates. It 

 establishes a strong presumption that 

 a similar relation would be found 

 among the men if we had the means 

 (which we do not) of determining the 

 ratio between all the married and un- 

 married men. 



The numerical representativeness of 

 our sample is also indicated by the fact 

 that the percentages of men and women 

 who filled out questionnaires are almost 

 the same as in the whole alumni body 

 of the period studied. The percentage 

 of men is 89.1 in the group return- 

 ing the blanks, and 87.1 in the Alumni 



Register, for the period 1870-1884. 

 The corresponding percentages of men 

 for 1885-1899 are 77.7 and 76.6. 



In conclusion, the sample studied has 

 doubtless been somewhat more pro- 

 lific on the average than all the gradu- 

 ates from 1870-1899, and then the 

 living group as a whole. It is repre- 

 sentative as far as concerns the ratios 

 between the sexes, and married and 

 unmarried women (probably men too). 



We are now ready for an analysis 

 of the facts in the light of the fore- 

 going criticisms. The general con- 

 clusions may be summed up in the 

 answers to three questions. 



a. Has the marriage rate signif- 

 cantly decreased? 



b. Has the birth rate diminished 

 significantly? 



c. Has the group produced enough 

 children to replace itself? 



a. The marriage rate has not sig- 

 nificantly decreased, either for the 

 men or the women. See Table 3. 

 Neither has there been any radical 

 change in the age at marriage. The 

 data for formulating the "Per cent 

 married" column for the men in Table 

 3, were derived from the question- 

 naires. The Alumni Register for 1921 

 does not indicate which men are mar- 

 ried. The Register does, however, dif- 

 ferentiate married from single women. 

 The data for the women in this column 

 were derived, therefore, from the 

 Register. About ninety per cent of the 

 men and fifty-six per cent of the 

 women are married. 



Doubtless the difference in the mar- 

 riage rates between men and women 

 is partly due to a desire on the part 

 of many of the women to enter upon 

 a career, rather than to accept oppor- 

 tunities for marriage. 



In only one case (between men of 

 the second and third decades) does 

 the difference between the percentages 

 of married men or women in succes- 

 sive decades exceed the probable error 

 of the difference. In that instance the 

 ratio of the difference to its probable 

 error is 1.07. It is evident that these 



