1950 was accidental or if there had been some definite trend in this direction 

 over time. It was possible to test this in two ways. Data were available 

 only by age categories for the 1930 population, hence all analysis was 

 limited to age categories rather than age-sex as in 1950. Chi Square was 

 again used as a measure of the relationships between the increasing and 

 decreasing towns for the periods 1930, 1940, and 1950. The values so 

 obtained were 55.7, 18.1, and 5.73, respectively, indicating that the popu- 

 lation distributions had been becoming more and more alike since 1930. 

 If some fixed distribution could be established and the age distributions 

 of the towns compared with this, such a measure might be more meaningful. 

 In 1943 a set of life tables by states and regions was published bv the 

 Metropolitan Life Insurance Company for the National Office of Vital 

 Statistics.* Thi.s included a life table for New Hampshire based on 1939-41 

 actuarial experience. This was revised to show the proportion that each 

 age group w^as of the total population. Each of the distributions was com- 

 pared against this standard. Table 3 shows the Chi Square values obtained. 



Table 3. Chi Square Values of Sample Population Distributions Compared with Population 

 Distribution from New Hampshire Life Table 1939-1941, Actuarial Experience 



Year Increasing Towns Decreasing Towns Difference 



1930 

 1940 

 1950 



The differences between these sets of values are, respectively, 85.3, 40.6, 

 and 8.9. In a 20-year period the difference in Chi Square value has de- 

 creased almost 90 percent. The population distributions have become more 

 alike. 



Sex Rario, 1940-1950 



The sex ratio (number of males per 100 females) in 1940 of the sample of 

 increasing towns was 99.8 and for the sample of decreasing towns 111.3 

 (t = 2.50; p = .027). In 1950 the ratios were 98.7 and 105, respectively, 

 (t m 1.22; p = .24). The change m sex ratio during the decade 1940 to 

 1950 bears out the earlier data developed through the Chi Square test. 

 The samples of towns are becoming more alike in population character- 

 istics. See Table 4. 



When the samples are divided into farm and non-farm components, the 

 sex ratios of the increasing sample are 107.7 and 94.3, respectively, while 

 the corresponding ratios for the decreasing towns are 101.4 farm and 107.9 

 non-farm. The value of "t" between the farm sex ratios is .81 (p = .43) 

 and between the non-farm sex ratios 2.06 (p = .055). 



Fertility Ratio, 1940-1950 



The fertility ratio (children 0-5 per 1000 women 15-44) for 1940 and 1950 

 was computed for each town (see Table 4) and the average ratio for each 

 set of towns computed. These were 554 for the increasing towns and 560 

 for the decreasing towns. The "t" test was applied to see if the difference in 



* State and Regional Life Tables 1939-41, Metropolitan Life Insurance Company, 

 National Office of Vital Statistics, Washington, D. C, 1948. 



