ABSTRACTThis paper presents a sample design for selecting streams in an investigation of the relation between water pollutian and water based recreation expenditures. The design is a proportional sample of clusters of streams stratified in a two‐way scheme. The clustering is achieved by sampling 71/2 minute topographical maps, and the stratification by using measures connected with water pollution to classify the clusters. The design is based on considerations of economy in data collection and computational efficiency. A distinctive feature of the design is that data collected are to be used in a regression analysis. Thus, the sample size is determined by trying to achieve a given degree of precision for the regression coefficient
展开▼
