classified Into specific code categories 

 were held aside for review by the coding 

 supervisor and project director. Specific 

 categories for the latter cases were 

 established when necessary. 



The work of all coders was checked by 

 the coding supervisor until em acceptable 

 level of coding consistency was achieved 

 both between and within coders. Tliere- 

 after a 10 percent check for the purpose 

 of maintaining this consistency level was 

 carried out. 



The punch cards were then prepsured 

 smd weighted as follows. In Detroit, the 

 Interviews completed in sampling xmits 

 selected frcm the white zone were dupli- 

 cated once since these interviewer assign- 

 ments had one -half the probability of being 

 Included in the sample as did the remaining 

 sampling units chosen for this survey in 

 that city. In addition the punch cards for 

 interviews completed in assignments which 

 had been subsampled were weighted according 

 to the subsampling rates. No atteii5)t was 

 made to substitute or weight for households 

 designated for the sample but not inter- 

 viewed. 



sample estimate to the result which would 

 be obtained from a complete census of the 

 population sampled, using the same ques- 

 tionnaire, interviews and interviewing 

 procedures . 



Practically all of the estimates 

 developed from the data collected in this 

 study are simple percentages of the re- 

 spondents having a pturticular opinion or 

 characteristic. In technical terminology, 

 these percentages are actually combined 

 strata ratio estimates, since the sample 

 design employed extensive geographic 

 stratification and. cluster sampling, in 

 which the number of respondents in each 

 cluster was subject to random sampling 

 vsuriation. "Rius, sanQjling errors were 

 computed using the formula for the 

 variance of a ratio estimate. 



This formula contains variance meas- 

 ures for the cluster average of both the 

 numerator and denominator of the computed 

 proportion or percentage estimate, as well 

 as a cov£u:iance measure for these two 

 averaiges. These measures were computed 

 from the average variance between clusters 

 within strata. 



The punch cards then received a 

 thorough error «uid consistency check on the 

 IBM Electronic Statistical machine. Where 

 necessary the punch cards were corrected 

 by reference to the specific questionnaires 

 corresponding to the cards in question. 



Sampling Errors 



The sampling error for a parti cvilar 

 estimate serves as a guide to the confi- 

 dence with \rtiich this estimate can be used. 

 It Is a measure of the closeness of the 



The chances are approximately 2 to 1 

 that the error, due to sampling, in a 

 parti culJtr estimate, will not exceed one 

 standard error; the chances are 19 to 1 

 eigainst a deviation as large as two stand- 

 ard errors from the result which would be 

 obtained with a coniplete census using the 

 same procedures. 



Estimates of the standard errors for 

 several items Included among the guided 

 association questions are shown in 

 Appendix Table 3. 



35 



