Seemingly Unrelated Multivariate Regression Estimators for Multi-Phase Sampling
Nadeem Shafique Butt, Shahid Kamal, Muhammad Qaiser Shahbaz
Multivariate estimators proposed by Ahmed, Hussin  for simultaneous estimation of multiple response variables (Yi) assumes that all response variables depend upon same set of auxiliary variables (Xi). This is not always feasible, in many situation different response variables may depend on different sets of predictors. In this paper new multivariate estimators for single, two and multiphase phase sampling have been proposed using the concept of Seemingly Unrelated Regression Estimation (SURE) and Mean square errors of the proposed estimators have been derived.