Table 2 



Sj\MPLING TOLERANCES FOR ESTIMATED NUMBER OF 

 ESTABLISm-EMTS APPEARING IN ANY PARI' OF SWflARY TABLES 



To sipvly the table to an estimate in a summary table, find the 

 line in column 1 of table 2 which is nearest to the number of establish- 

 ments in the estimate for which sampling tolerance ia desired. For any 

 estimate except one imrolving a breakdoim by a particular type of 

 establishment obtain the approximate sampling error from the second column 

 in the table. V/hen the estimate is for a particvdar type of establishment 

 such as restaurant, or drug or proprietary store with fountain service, 

 use the third or fourth column depending on the type of establishment. 



As an example of the use of the table consider the estimates 

 of the number of public eating places obtaining their fish and shellfish 

 supplies from wholes al.ers (table 55). In the Northeast Region an 

 estimated 1^5,200 establishments of all kinds, received their supplies from 

 wholesalers. Turning to column 1 of the table it is foiuid that 1^5,200 

 is roughly 5/6 of the way through the range between the figures 20,000 

 and 50,000 shown in that column. By interpolation, the sampling error 

 figure for the estimate U5,200 is 5/6 of the distance between 2,i400 and 

 i|,000 in column 2 or roughly 3,700. 



This sampling error of 3,700 means that 2 times out of 3 the 

 true figure of all establishments in the Northeast Region which purchase 

 their fish and shellfish supplies from wholesalers will be within plus or 

 minus 3,700 establishments of the estimate of 1(5,200, The true figure would 

 be the figure obtained by taking a complete census of all the public eating 

 place establishments in the Northeast Region, 



11 



