COST OF LIVING INDEXES COMPARED 123 
covered. The National Industrial Conference Board number 
is based on a varying number of cities for the different major 
items. The average since 1920 has been 170 for shelter; 79 
for clothing; 59 for coal; for gas and electricity in 1925, 143 
cities; for carfare, 288 cities. In December, 1925, these 
aggregated a population, based on the 1920 census, of 33,105,- 
454 for rents; 27,641,433, for clothing; 25,917,061 for fuel; 
32,695,073 for gas and electricity, and 37,770,114 for carfare. 
Changes in the cost of other sundries items, such as maga- 
zines, insurance, candy or tobacco, while based on a few 
reports, are of application to the country as a whole. 
This number of cities has not always been included, how- 
ever, in either series.! Through December, 1917, the Bureau 
of Labor Statistics index is based on 18 shipbuilding centers; 
in December, 1918, 13 more industrial cities were added; 
in December, 1920, Washington, D. C., was included. The 
National Industrial Conference Board practically completed 
its basis in 1920, also. The places included differ consider- 
ably as between the two series, but both include the largest 
cities. The average population covered by the National In- 
dustrial Conference Board reports is several millions greater 
than that covered by the Bureau of Labor Statistics. 
In some respects, the population coverage is not so im- 
portant in determining the adequacy of the sample as other 
considerations. Geographical location of the cities appar- 
>ntly affects changes in the cost of coal and to some extent 
changes in the cost of food; local conditions are more exclu- 
sively responsible for rent changes. Clothing price changes, 
on the other hand, are more nearly a matter of individual buy- 
ing and selling policies than of geographical location,” but 
local industrial conditions in some instances seem to have 
played an important part in the increase or decrease of prices. 
The purposes of the two series should be kept in mind in 
any comparison of their coverage. The Bureau of Labor 
Statistics is interested primarily in ascertaining changes in 
cost in the individual cities and secondarily puts these 
1 Except that for food, prices were used for all of the cities in which they were 
secured. 
21t is of interest that, according to the Bureau of Labor Statistics series, the 
cost of clothing increased more in New York, the clothing center of the country, 
-han in anv other single locality. See also footnote 1, p. 119, of the present volume.