Difference-Cum-Exponential-type estimators for estimation of finite population mean in survey sampling

Area of Science:

• Statistics
• Survey Methodology
• Data Analysis

Background:

• Conventional measures for population mean estimation using bivariate auxiliary information are sensitive to outliers.
• Non-conventional measures (e.g., quartile deviation, tri-mean) offer robust alternatives but are underutilized in survey practice.
• There is a need for improved estimation techniques that are resilient to extreme values in auxiliary data.

Purpose of the Study:

• To propose novel difference-cum-exponential-type estimators for population mean estimation.
• To utilize bivariate auxiliary information based on non-conventional measures.
• To evaluate the performance of these estimators under simple and stratified random sampling.

Main Methods:

• Development of difference-cum-exponential-type estimators incorporating non-conventional auxiliary measures.
• Derivation of mathematical properties, including bias and mean squared error.
• Empirical validation using real-life datasets to compare with existing estimators.

Main Results:

• The proposed estimators demonstrate superior performance compared to conventional estimators, particularly in the presence of outliers.
• Non-conventional measures significantly enhance the efficiency and robustness of population mean estimation.
• Theoretical derivations of bias and MSE support the practical findings.

Conclusions:

• The suggested difference-cum-exponential-type estimators utilizing non-conventional measures are effective for population mean estimation.
• These estimators provide a robust and efficient alternative to traditional methods, especially in datasets with extreme values.
• The study advocates for the adoption of non-conventional measures in survey sampling to improve parameter estimation.