A Power Transformation Ratio-type Estimator of Finite Population Mean in Stratified Random Sampling
Authors
-
IkughurJ. A.
Department of Statistics, Joseph Sarwuan Tarka University, Makurdi, P.M.B. 2373, Benue State, Nigeria
-
Uba T.
Department of Statistics, Joseph Sarwuan Tarka University, Makurdi, P.M.B. 2373, Benue State, Nigeria
-
Adamu R.A.
Department of Statistics, Federal Polytechnic, Wannune, Tarka, P.M.B. 102355, Benue State, Nigeria
Abstract
In this paper, a ratio-type estimator of finite population mean in stratifiedrnrandom sampling based on the Srivastava(1967) estimator have beenrntheoretically and empirically studied. Theoretically, the bias and variancernfor the separate and combined estimators were obtained and compared itrnwith the bias and variance of the traditional separate and combined ratiornestimators. Other criteria used to determine the most efficient estimators are:rnstandard Error (SE), Coefficient of Variation (CV) and Percentage RelativernEfficiency (PRE). Comparison of the estimators empirically, shows that thernmodified estimators are more efficient than the traditional separate andrncombined ratio estimators in all conditions. In addition, varying values ofrn indicates that even if deviates from its exact optimum values thernmodified estimators s y and c y will be better estimators than the usualrnestimators RS RC y, y , y .Based on such findings, we suggest the use of thernproposed estimator in stratified random sampling as better estimators overrnthe traditional ratio estimators in stratified random sampling for practicalrnsituations.
Keywords: Power transformation, Population mean, Stratified random sampling, Auxiliary variable.