To maintain the statistical properties of the expanded estimates, the telephone 

 sampling effort was probabalistically allocated at the household level. Interview allocations 

 for each county were based on the square root of the population within each county. _\J Jhe 

 allocation of calls made in each telephone prefix was based on the frequency of households 

 assigned that prefix. The appropriate number of sample points (household telephone numbers) 

 was then randomly generated for each county for each wave without duplication. 



Telephone interviews were conducted between 8:30 am and 9:00 pm local time. Up to 

 three attempts were made to reach each household and up to six calls were made to households 

 determined to contain marine fishermen to complete the questionnaire. Spanish-language 

 interviews were conducted as required. Information on marine recreational fishing activity 

 was solicited from each fishermen in the household or from a responsible adult when 

 appropriate. A sample telephone questionnaire is included as Appendix A. 



INTERCEPT SURVEY METHODS 



The intercept portion of the survey consisted of on-site interviews which gathered 

 catch and demographic data from marine recreational fishermen in three modes: 

 party/charter boat, private/rental boat, and fishing from shore (e.g., man-made structures, 

 beaches and banks). This differed from previous surveys when estimates for beach/bank 

 fishing and fishing from man-made structures were tnade separately. However, raw intercept 

 and telephone data files are maintained with the original mode of fishing variable intact. 

 Sampling was conducted continuously in six 2-month sampling periods from January through 

 December 1986, with the exception of the Atlantic coast north of Georgia, where January and 

 February were not sampled. 



The allocation of interviews by mode and wave was based on empirical data and 

 informed estimates from the 1979-1985 surveys. The sample allocation among the various 

 modes was modified as sampling results from 1986 became available. Complete coastwide site 

 lists were created and sites were randomly selected to meet the mode allocations. Sampling 

 was scheduled to include all weekdays, weekends and holidays. 



Interviewers were assigned to randomly selected sites where fishermen were 

 interviewed at the completion of their fishing trip. At heavy use sites, samples were taken of 

 every n^^ fisherman at the completion of the angler's trip. The interview consisted of an 



— U.S. Department of Commerce, Selected housing characteristics by States and counties: 

 1980. Publication Number HC 80-51-1 (October !98r). 



