2. Using the above stratification and 1950 census statistics and 

 supplementary information as indicated belot^, a sample of 668 blocks was 

 selected with known probabilities using multi-stage sampling procedures. 

 The primary sampling units (p.s.u.'s) were selected mthin each stratum 

 with probabilities proportionate to size (p.p.s.)- These p.s.u.-s con- 

 sisted of whole SMA'Sj individual cities or towns in the urban city size, 

 and rural portions of non-SM counties. Sampling xdthin each selected 

 p.s.u. was as described in the following paragraphs. 



3» In all p.s.u.'s, a nrLnimTjn of foxir blocks was selected in each 

 selected city, town, or minor civil division (m.c.d.). Measures of 

 block size were provided by Market Research Corporation of America (MRCA) 

 for each selected block. By vising the intra-block sampling fractions, a 

 three-state sanqile of households was selected with equal probabilities, 

 and hence a self-weighting sample of households vas obtained. 



U. In the block-city portion of the SJiA's, block cities were 

 selected ^dth probabilities proportionate to size. A listing of these 

 selected block cities, together x^ith their measures of size, was pro- 

 vided by Market Research Corporation. iVithin each selected block city, 

 blocks were selected with probabilities proportionate to size, using 

 systematic sampling x-iith a random start. A miniinum block size of ten 

 census households was established, combining small blocks (including 

 zero blocks) vrith adjacent larger ones in an unbiased manner. A listing 

 of selected blocks by tract and block number, together xdth their 

 respective measures of size, and one copy of a reproduction of the 

 census block map with the selected blocks indicated thereon were pro- 

 vided by Market Research Corporation. 



5. In the non-block area portion of the SMA's, minor civil divi- 

 sions (m.c.d. 's) were selected vjith p.p.s. A list of selected m.c.d. 's 

 was provided together with measures of size. Within each selected m.c.d., 

 four blocks (or small areas) were selected with p.p.s. where co\int units 

 were available on county maps and with equal probabilities where they 

 were not. One copy of a map or map reproduction was provided by Market 

 Research Corporation for each selected m.c.d. on which the selected 

 blocks were indicated. 



6. In the urban city size, cities or tovms were selected x^ath 

 p.p.s., and a listing of selected cities or towns was provided together 

 idth their measures of size. Within each selected cit2' or toTrm, four 

 blocks were selected vjith equal probabilities. A nap or map roprodnc- 

 tion Tjas provided for each selected town or city on which the selected 

 blocks were indicated. 



7. In the rural city size, within each selected county, a m.c.d. 

 was selected with p.p.s. and with replacement for each group of four 

 blocks, A listing of selected m.c.d. 's together with their measures of 

 size, xi/as provided. Within each selected m.c.d., four blocks or small 

 areas were selected with p.p.s. where count units were available on 

 county maps and tdth equal probabilities where they were not. A map of 

 selected m.c.d. «s was provided on t'jhich was indicated the selected 

 blocks or areas. 



