THE ALLEGHENY COLLEGE BIRTH RATE 



H. R. Hunt 

 University of Mississippi 



( Coiiliiiucd from the Max Niiiiihcr) 



THE difference in family size per 

 married alumnus in the two 

 periods is .29 - .20 children. 

 A chance difference of this magni- 

 tude will occur 33 times out of 

 every 100. The difference is there- 

 fore not statistically significant ; it 

 may have been due to chance. It 

 should be recalled, however, that the 

 living graduates in the period 1870- 

 1884 are a more highly selected group 

 from the standpoint of length of life 

 than those in the period 1885-1899. 

 Therefore the average fecundity in the 

 earlier group may have been higher 

 than that in the later one, because the 

 long-lived men have the larger fami- 

 lies. Had both groups been equally 

 selected for longevity, the difference in 

 family size might have been still less. 

 The difference between the two 

 periods in the average number of 

 children per male graduate is .30 - .19. 

 This difference is likewise not signifi- 

 cant statistically. But married alumni 

 with incompletely reported families 

 have been excluded in Table 5. If the 

 number of unmarried individuals in 

 this computation is reduced so that the 

 percentage of unmarried is the same as 

 in the whole data derived from the 

 questionnaire (S% and 11% for the 

 periods 1870-1884, 1885-1899, respect- 

 ively), then the two averages will be 

 2.69 (1870-1884) and 2^36 (1885- 

 1899). The percentage of married in- 

 dividuals among the women graduates 

 reported in the questionnaires is not 

 far, it will be recalled, from the per- 

 centage among all the women gradu- 

 ates as reported in the Alumni Regis- 

 ter. This was true whether reporting 

 women with incompletely reported 

 families be excluded or included in 

 making the computation. In other 

 words, the questionnaire provided a 



fair sample of the women as far as the 

 proportions of married and unmarried 

 persons were concerned. Our sample 

 of men is therefore probably repre- 

 sentative in the same respect. All 

 available data, therefore, indicate that 

 the numbers of unmarried men we 

 have used in computing the "Children 

 per male graduate" in Table 5 are not 

 far from what they would have been 

 had accurate determinations of percent- 

 ages of married men in the whole 

 body of graduates been available for 

 use. 



The data shown in Table 5 do not 

 meet all the requirements of a rigor- 

 ous analysis. If the wife is now dead 

 or divorced, the alumnus may marry 

 again and have more children by the 

 second wife. In cases where the wife 

 is less than forty-five years of age, the 

 family may yet increase in size be- 

 cause the mother is still capable of 

 bearing children. The averages in 

 Table 6 are based on families in which 

 (1) the husband (a graduate) is dead 

 (13 cases), and (2) families in which 

 both husband (a graduate) and wife 

 are living, but the wife is not divorced 

 and is 45 years old or older. 



The difference between the mean 

 number of children of married alumni 

 in the two periods is .24 - .22. This 

 difference is not statistically signifi- 

 cant. It may have been due to chance. 

 Had the two groups been subjected for 

 equal lengths of time to lethal selec- 

 tion, the difference might have been 

 even less, as pointed out for Table 5. 



The divergence between the average 

 numbers of children per male gradu- 

 ate, .29 - .22, is likewise not statistical- 

 ly significant. The percentages of un- 

 married persons used in making these 

 computations are the same as in the 

 whole data obtained in the question- 



139 



