Probably the population similarity between the increasing and decreas- 

 ing towns cannot be achieved quickly because of the relative slowness of 

 population change once migration slows down. The degree of similarity 

 will be affected by the relative values of the birth and death rates from 

 period to period. The trend, however, from 1930-1950 will be toward greater 

 similarity. 



There has been a general assumption that economic structure affects popu- 

 lation composition. The test of the belief that they are independent of one 

 another is only possible in an area of long and stable settlement. It is 

 known that areas possessing many new migrants differ in population com- 

 position from older settlements. Independence is testable only in areas 

 where resource utilization is relatively stable. 



If a general type of family structure is common within our culture, how- 

 ever, once migration slows down the composition of the populations of in- 

 creasing and decreasing towns ought to become more alike despite differ- 

 ences in economic structure. 



The fact that some towns are still increasing and others decreasing indi- 

 cates that adjustment to resources is not perfect. It would then follow that 

 the resources in the increasing towns will support more people while those 

 in decreasing towns will support fewer. The values of economic factors 

 ought then to be somewhat larger in the increasing towns. Since the migra- 

 tion is relatively light, these differences ought also to be minor. 



This sort of analysis differs from the more usual treatment of the popu- 

 lation in terms of rural farm, rural non-farm, and urban. This different 

 treatment was used because of the belief that these census residential cate- 

 gories in New England have little meaning with respect to spatial char- 

 acteristics. The Minor Civil Divisions of New England are likely to in- 

 clude elements of all these in varying proportions to one another. For this 

 reason there is the possibility that conclusions drawn from population 

 analysis based on census residential categories may be misleading when 

 applied to New England towns. 



Procedure 



For a number of reasons it was evident that some sort of sampling pro- 

 cedure would have to be employed. First, the changes in population for 

 all towns in the State, either increase or decrease, ranged from percent 

 to 50 percent. The bulk of the towns showed percent to 8 percent change. 

 Any attempt to dichotomize into "gainers" and "losers" then would in- 

 clude in each category a substantial number that were stable. Second, the 

 time demands for the analysis of all the Minor Civil Divisions in New 

 Hampshire (248) were impossible to meet with the research resources 

 available. And third, it was desired to eliminate as far as possible the 

 short term fluctuations which occur in the smaller towns when a single 

 mill closes, then re-opens a year or so later. Such towns face a different 

 kind of problem (that is, fluctuation of population) rather than general 

 increase or decrease. There was the possibility that the characteristics 

 of such a town might differ markedly from one which had a long history 

 of steady increase or decrease. 



Sample Selection 



The first step in the selection of the sample was to set up criteria which 

 would yield towns having different directions of population change and 

 adjustment. The standards finally decided upon were: 



