Comparison of Some Almost Unbiased Ratio Estimators

A survey sampler is always desperate for designing a best estimator and devotes endless effort to achieve that. As a result a reasonable number of estimators developed especially for the most urged parameter - Average. A notable development occurred in designing such estimators by exploiting the ancillary information at hand. It recognized several non linear estimators, which are analyzed and compared in terms of biases and mean square errors using linearization techniques, in which the higher order terms are sacrificed. So, it is worth investigating their actual performance before using these. The present paper considered some unbiased ratio-type estimators (URE) and their performance were accessed by simulating under bootstrap method. It is recommended that the Beale's (1962) estimator as a reasonable one for estimating population mean under the assumption of bivariate normal law for the universe.