On Abid Et Al (2016) Ratio Estimators for Finite Population Mean Using Non-Conventional Location Parameter

Abstract:

Ratio estimator is most effective for estimating population mean when there is a linear relationship between study variable and auxiliary variable when have positive correlation. Abid et al proposed some new modified ratio estimators in simple random sampling, and improved ratio estimator. The first six ratio estimators of Abid et al (2016) for finite population mean in simple random sampling using Tri-mean and mid-range with correlation coefficient and coefficient of variation as supplementary information was used in this work. The aim is to use data simulated from some distributions with varying sample sizes to determine which of the six estimators is more efficient than others. Friedman test was used to rank the bias and the mean square error (MSE) of each of the distribution with varying sample sizes. The sixth estimator have the minimum bias and minimum MSE than other estimators in seven distributions out of the eight distributions computed and that made it to be considered as the best estimator. Only one distribution is odd among the eight distributions and it has the third estimator as the best.