Abstract
A selection procedure for selecting an inclusion probability proportional to sample of size (2 ⩽ n < N) is proposed. The simplicity in sample selection and in computing inclusion probability for a pair of units in the sample is achieved by using combinatorial properties of partially balanced incomplete block design. A procedure has been illustrated by an example and also comparison of Horvitz-Thompson and Sen-Yates-Grundy estimators has been discussed.
Keywords
Get full access to this article
View all access options for this article.