reproductive behavior was significant statistically. The value of "t" was 

 .15 (p = .87) indicating no significance in the difference. 



Farm — Non-farm Fertility 



The populations were then divided into farm and non-farm components 

 and the comparisons again made between the average fertility ratio of 

 these segments of the population of the increasing and decreasing towns. 

 The rates were found to be 580 and 521 for the farm populations (t = .48; 

 p = .64) and 554 and 594 (t ^ .63; p = .54) for the non-farm popu- 

 lations, respectively. 



On an unstandardized basis the non-farm population of the increasing 

 towns is more prolific than the non-farm populations of the decreasing 

 towns.* The farm population of the decreasing towns is more prolific than 

 the farm population of the increasing towns. The differences are of approxi- 

 mately the same magnitude and explain the very slight differences noted 

 when total populations are compared. 



The important thing to be noted is that the populations are comple- 

 mentary in terms of their reproductive behavior. As far as population 

 growth or decline of the area is concerned, the farm and non-farm sub- 

 populations should be considered as a unit. Differences in age distribution 

 of the female population of the farm and non-farm segments of the in- 

 creasing and decreasing towns might be responsible for their reversal of 

 reproductive behavior. The samples were divided accordingly and the 

 average number of women 15-44 was compared using the "t" test. The 

 differences were not significant. The differences in reproductive behavior 

 apparently are due to social factors, not to an excess or deficiency of women 

 in the reproductive years of life in either group. 



The average fertility ratio in 1940 for the increasing towns was 403 and 

 for the decreasing towns 401 (t = .05; p = .98). Since 1940 both sets 

 of towns have followed the national trend in fertility, but the increasing 

 towns increased their fertility rates more noticeably than the decreasing 

 towns. Since we know that the age-sex composition of these sets of towns 

 in 1950 is very much alike, it probably is necessary to attempt to explain 

 this difference in rate of change in social terms. Previous studies have 

 shown that the number of children per farm family bears an inverse re- 

 lationship to income. f A recent study from North Dakota indicates that 

 levels of living tend to rise in those counties experiencing greatest loss of 

 population.:]: It may be that some such trend operates in New Hampshire 

 and would explain the failure of the decreasing towns to increase their 

 fertility rates at the same rate that the increasing towns were increasing 

 theirs. See Table 2. 



Dependency Ratio 



One of the measures used extensively in inferring the relative burden of 

 supporting social institutions is the dependency ratio. In this study the 



* The similarity in number of females 15-44 was sufficiently close to make the 

 work of equating to equal numbers of women unnecessary. 



t Loomis, C. P. and Beegle, J. A., Rural Social Systems, pp. 100-105, Prentice Hall, 

 New York, 1950. 



t Anderson, A. H., "Changes in Farm Population and Rural Life in Four North 

 Dakota Counties", A.E.S. Bull. 375, North Dakota Agricultural College, Fargo, N. D., 

 April, 1952. 



11 



