8. The sample allocation was approximately as outlined in appendix 

 table 2. 



9. The households vjithin the blocks x^ere selected as follows: 



a. From the measures of block size provided by ERCA under 

 3 above, a sampling rate was calculated so as to give an 

 expected total number of households of 2,001 or an average 

 number of three households per block. 



b. A cruise by car T^ras made in each block to count the 

 number of households. This number was multiplied by the 

 sampling rate to give a calculated number of households to 

 sample. If the calculated number of households in a block 

 was six or less, this was the sample number. 



c. If the calculated number was greater than six, only six 

 households were sampled. The total number of households in 

 a block was divided by the sample nwaber to give a sampling 

 interval for the block. Numbering the households along a 

 pre-designated route, a random start plus the sampling 

 interval vras used to select the sample households. 



d. To give a self -weighting sample, the sample households 

 were duplicated sufficiently to arrive at the calculated 

 number of households for each block. liJhere only a portion 

 of the sample households needed to be duplicated, thase were 

 selected randomly. 



A total of 1,883 interviews was completed. Unsuccessful attempts 

 were made to contact 129 additional not-at-home households. As the 

 sample was drawn proportionate to population, a high percent of the 

 interviews was taken in the coastal areas where the population of these 

 states is concentrated. 



In duplicating selected records to^ive a self -weighting sample 

 (as described in 9d above), 221 additional questionnaires were added to 

 the study, for a total of 2,06U. Throughout the study, reference to the 

 number of households includes the duplicated questionnaires. Of the 

 2,06U households included in the analysis, l,^hh were from California, 

 337 were from Washington, and 183 from Oregon, 



The field survey work in Oregon and in the rural and nonblock areas 

 of VJashington and California was completed by Oregon State Coj2ege 

 personnel. In Seattle, San Francisco, and Los Ange3.es professional 

 interviewers viere hired to do a substantial part of the field work, 

 under the supervision of a staff member from Oregon State College. 



I 



lit2 



