New Difference Estimator in Two-phase Sampling using Arbitrary Probabilities

A new difference estimator has been constructed in two-phase sampling using two auxiliary variables w and x.The first phase sampling unit has been selected with probability proportional to measure of size and second phase sample is selected with equal probability without replacement. The proposed estimator has been found to be more efficient as compared to Raj (1965) in which single auxiliary characteristic has been used. The variance expression can be easily extended to p-auxiliary variates.