Estimation of the exponential memory-type ratio estimator using simple random sampling

Abstract

It is significant in survey sampling that the population mean is estimated with the highest possible accuracy. A new memory-type exponential-ratio estimator is defined and its performance under simple random sampling (SRS) is evaluated in this research. The proposed estimator is shown to utilize past auxiliary information very effectively which results in a significantly lower mean square error (MSE) than other existing estimators. Based on theoretical derivations and genuine datasets, the proposed estimator was tested, and it was found to continually outperform traditional ratio-type estimators. Empirical results, including those from the comparisons of precision relative efficiency (PRE), confirm its superiority using reduced variance and increased efficiency. The results obtained indicate that the use of memory type modifications in exponential-ratio estimators enables the attainment of more accurate estimates of the population mean in sample surveys.