The United States was divided into 16 parts by 4 separations, 

 each in 2 directions. The first separation was made by the 

 density of the popnlation: 



1. Rural Americii 



2. Small-town America 



3. Big-town and suburban America 



4. Large-city America 



Inherent inclination to hunt and fish provided the second four- 

 way cut: 



a. Heavy hunting-fishing America 



b. Better-than-average hunting-fishing America 



c. Average hunting-fishing America 



d. Below-average hunting-fishing America 



The United Stat(>s was sliced into 16 parts so that a separate 

 suhsample could be taken from each part and the results added 

 together. This served to spread out the sample, thereby pro- 

 viding better dispersion and representativeness. It also made 

 possible the differing sampling rates which enhanced the sample's 

 efficiency. 



This division was decidt»d upon because it provided the greatest 

 efficiency under two premises: 



1. The more urbanized an area, tiie less its inhabitants indulge 

 in hunting and fishing. 



2. All other things being equal, the inhabitants of some States 

 have a stronger inclination to hunt and fish than those of other 

 States (because of opportunity, customs, and other influences). 



To test these premises and measure their importance, a correla- 

 tion analysis was made of the number of resident hunting and 

 fisliing licenses issued by each Stale against tiiat State's urbaniza- 

 tion. This is a statistical procedure that measures and describes 

 mathematically the relation between two sets of facts. The rela- 

 tion proved to be quite strong in this case (36 percent). The 

 strength of the relation showed that it would be most efficient to 

 sample the large cities, big towns, small towns, and rural areas at 

 different rates. 



The statistical formulas did all the work in first picking com- 

 munities, then neighborhoods, and then houses. A "community" 

 was either a center of population concentration, from the largest 

 city down to the smallest village, or what was left over in a 

 township or county when all of these centers were removed. 



Communities and neighborhoods for the survey were drawn 

 from Census Bureau lists by means of statistical formulas. The 

 number of houses to be included in each neighborhood had been 

 previously determined. It varied among the 16 parts of the 

 United States, but it was the same in all communities sampled in 

 any one of the parts. The number of houses was based in large 

 part on the number of hunters and fishermen expected to be present 

 in each particular part. 



While the number in each neighborhood was thus predetermined, 

 the selection of the actual sample houses was again a matter of 

 chance selection by statistical formula. The interviewer listed all 

 homes in the neighborhood in a certain order on a special listing 

 form but sampled only those whose listings fell in previously 

 marked lines on the form. 



THE INTERVIEWING 



The survey was conducted on a house-to-house basis b\' trained 

 interviewers of Crossley, S-D Surveys, Inc. Before the actual 

 survey, the elaborate questioimaires, one for fishing and one for 

 hunting, were pretested in a number of areas scattered throughout 

 the United States. Field interviewing for the survey began on 

 January 7, 1956, and continued for appro.ximately 2 months. 



A program of advance publicity was carried out before the start 

 of field interviewing. This included official releases to news- 

 papers, magazines, and radio and television stations by the Fish 

 and Wildlife Service, and a number of radio appearances bv mem- 

 bers of the Crossley firm, all inviting public cooperation. 



Field supervisors of the Crossley firm attended training sessions 

 on the survey in the New York office and in turn held similar 

 training sessions for interviewers in their areas. Each interviewer 



47 



