This study aims to explore the preference order regarding the utilization of different estimation methods in sample surveys. Through empirical analysis, the research examines both the conventional simple random sampling without replacement estimator and the efficiency of double sampling for ratio and regression estimators. The objective is to identify the methods that the estimator is the most efficient. Double sampling procedure was adopted, and comparing the minimum variances empirically which was used to obtain the most efficient estimator using the data collected from the variable of interest (expenditure) and the auxiliary variable (salary). In the first phase, a sample size of (150, 120, 80, and 60) was chosen from the population and in the second phase a subsample of size (70, 55, 45, and 30) was selected from the first phase, each at four different levels (1, 2, 3 and 4) without replacement. Of the three sampling methods, namely double sampling for ratio estimator, simple random sample without replacement, and double sampling for regression estimator, the last one shows the least variability, making it the most effective estimator in terms of efficiency. Consequently, when the auxiliary variable is accessible, it is advisable to utilize the double sampling for regression method in order to enhance the accuracy of estimating the population parameter.

Keywords : Double sampling, ratio estimator, regression estimator, simple random sampling without Replacement, minimum variances.