The error or sampling tolerance figures in table 2 may be 

 doubled and used with even less risk. For exanqjle, vrtien desired, the 

 figure of 3,700 in the above example may be doubled to 7,liOO. In this 

 event it raay be stated that 19 times out of 20 the true figvire of all 

 establishments in the Northeast Region which purchase their fish and 

 shellfish supplies from wholesalers will be within plus or minus 

 7,U00 establishments of the estimate of U5,200. 



The analysis of the information obtained from the firms which 

 participated in the survey is given in the foHovfing sections. For each 

 topic covered in the survey there is usually shown in the anali'sis a 

 breakdown of the United States total response by regions, by type of 

 establishment, by city size groups and by sales size groups. The latter 

 analysis refers solely to the sales size based on receipts from sales of 

 food. For example, a large downtoxm drug store with sales of $2^0,000 

 or more per annum with sales in its lunch counter section of $50,000 

 per annum would be classified in the $i|0,000 to $99,999 sales size group. 

 Whenever fewer than the above four types of breakdown of the data are 

 shown it was either impossible or ingjractical to prepare the type of 

 breakdown vdiich is not shown. 



12 



