Robustness of modified factor-type estimators under super-population model

Abstract

In this paper, efficiency and robustness of modified factor-type estimators have been study under super-population model in relation to other existing related estimators of population mean. Bias and Mean Squared Error (MSE) of suggested estimators under non-response model have been obtained and empirical study was done using Data 1 and 2. The robustness of the considered estimators were obtained by averaging their expected MSE and ranked accordingly to their level of efficiency. The results revealed that the efficiency of all the estimators considered increases and decreases as the relationship (ρ_xy) between the study and auxiliary variables approaches linearity when ρ_xy > 0 and diverges from linearity when ρ_xy < 0 respectively. The results also revealed that the modified estimators compete favourably with almost the estimators considered in the study.