The sample in each kind of business was allocated to the 

 components described above ±a approximate proportion to their ingwrtance 

 in total sales. 



Eating Places 



All establishments (group I stores) which the Census Bureau 

 personally enumerated in land segments during December 19Bh, and January 

 and Febmaiy 1955 were selected. From lists of large stores located in 

 primary sampling units hlO establishments were chosen with probabilities 

 proportionate to their weighted sales-size in January 1955. From the 

 lists of group II stores a sample of 2liO establishments was drawn \jy 

 first selecting organizations with probabilities proportionate to the 

 organization sales in January 1955 and then selecting establishments 

 from the organizations with equal probability. 



Drinking Places and Drug and Proprietary Stores with Fountain 



This sample was dravm in the same manner as for eating places 

 with the following modifications. Only one month* s panel of land segments 

 (February) was used and from the lists of large stores in prijnary sampling 

 units 20 drinking places and 70 drug and proprietary stores with fountain 

 were chosen. From the lists of group II organizations one drinking place 

 and 70 drug and proprietary stores with fountain were selected. 



The sample of hotels was selected from a list compiled in 

 connection with an annual survey taken for 1953 which employed the Monthly 

 Retail Trade sample design. From this list hMh hotels were selected with 

 probabilities proportionate to their weighted 1953 armual sales of meals 

 and alcoholic beverages. 



81 



