Mean Estimation in Time-Scaled Surveys: a Dual Auxiliary Information Approach with Applications to Agricultural Data

Accepted March 2026

Abstract

Mean estimate in sampling is essential for quickly inferring population characteristics when a complete census is unfeasible. The incorporation of dual information results in an increase in the accuracy of the estimations used. This study investigates the effectiveness of the exponentially weighted moving average in enhancing estimation accuracy. It introduces a memory-based exponential-cum-product estimator that aims to better capture temporal relationships. This is achieved by utilizing supplementary information derived from dual-transformed auxiliary variables under the simple random sampling without replacement framework. The bias and mean squared error of the proposed estimator are derived up to the first order of approximation. Furthermore, the study includes detailed empirical analysis with application to recent agricultural data and extensive simulation studies, which serve to compare the performance of the proposed estimator with that of other existing methods.