The detailed findings are presented in the fornn of percentaged 

 distributions to the responses to various questions asked. 

 While the percentages are based on the weighted number of 

 tabulation cases, the number appearing in parentheses at the 

 head of each column is the actual number of interviews con- 

 ducted. 



Throughout the detailed tabulations a single asterisk ( * ) has 

 been used to denote that the percentages may add to more than 

 the total or subtotal since some respondents give more than 

 one reply to the question. 



Any unusual circumstances relating to the Detailed Findings 

 are explained by footnotes on the tables to v^ich they apply. 



With only a few exceptions all of the detailed tables are shown 

 in the same sequence for all ten cities. The exceptions are 

 several cases where the number of respondents was so low 

 that the table had no statistical significance and was therefore 

 not shown. 



Tables a^ through i_are a statistical description of the kinds of 

 establishments included in each city sample. 



m. RELIABILITY OF STUDY RESULTS 



A. Sampling Error 



Standard Error of the Proportions 



For all percentages, the standard error may be estimated by 

 using the formula for simple random sampling (the symbol 

 "d" denotes standard deviation). 



dp - 



v/f 



and making appropriate adjustnnents for estimated loss of 

 statistical efficiency of the sample due to clustering and for 

 the gain due to stratification and the finite population factor. 



11 



