Ratio estimator for double sampling procedure with non-response: An empirical study

Abstract

This study proposes a ratio-type estimator for population mean estimation using auxiliary variables with double sampling in the presence of non-response. The study provides expressions for the constant, bias, and mean square errors (MSE) of the proposed estimator and compares it with ten existing estimators. The study employed the secondary source of data collection to evaluate the efficiency of the proposed and existing estimators by analyzing five natural populations from three different sources. The performance of ten (10) estimators was considered in this study. The findings suggest that the proposed estimator and the H estimator provide more accurate and precise estimates of the population mean using an auxiliary variable. Additionally, the study found significant differences amongst the mean values of the constant and bias for the different estimators. A Dunn Kruskal-Wallis multiple comparison tests with the Bonferroni method was performed to ascertain the pair of estimators that contributed to the significant difference observed. When estimating the population means using an auxiliary variable, the proposed estimator outperformed other existing estimators that were taken into consideration in the study