PAKISTAN JOURNAL OF STATISTICS, vol.24, no.3, pp.167-172, 2008 (SCI-Expanded)
We propose ratio estimators that can be used under unequal probability designs by adapting Horvitz-Thompson estimators to ratio estimators in literature. Mean square error (MSE) of all proposed ratio estimators is obtained and compared with the MSE of classical estimators. By these comparisons the conditions, which make each proposed estimator more efficient than the classical one, are found. The theoretical results are supported by a numerical example.