Use of Double Sampling in Repeated Surveys

Abstract

In the present study sampling on two occasions has been extended to cover the case of double sampling for regression and stratification. The particular are of double sampling for stratification when there are only two strata, one consisting of only zero units has been investigated in greater details. The case of sampling with varying probabilities of selection has also been included. The optimum solution for n2, n and q has been obtained by the method of successive approximation. The behavior of the efficiency of double sampling for variation of p1 and 2 has been examined separately. It has been founded that for p1 small, double sampling is highly efficient then simple random sampling. For p1 very high double sampling is not very efficient and if it is applied at all, the size of preliminary sample should be kept as small as possible. Similarly, the efficiency of double sampling is very high compared to simple random sampling of 2 is small and decrease first rapidly and then gradually. After a certain limit, the fall in efficiency is negligible even if increase in 2 is significant.