Regression estimator based on median for population mean estimation under non-response analysis

One of the challenging components of survey sampling over the past few decades has been estimating the mean through a median of the study variable and significant effort has gone into developing effective estimators. In this article, we suggest a median-based linear regression type estimator that takes into account the mean of the auxiliary variable under non-response, as well as the median and mean of the study variables. In the first situation, it occurs only on the study variable, and in the second situation, non-response occurs mutually in the study and the secondary variables. We are driven towards it because, despite its infrequency, median-based estimate using the secondary variable under non-response has not been explored. Under the first-order approximation, equations for the biases and MSEs of the suggested and existing estimators are obtained. The empirical study compares the proposed and current estimators using a real data set. Based on actual data sets, our proposed estimator yields a lower mean square error (MSE) and a greater percentage relative efficiency when compared to existing estimators. Both theoretical and numerical data show that the proposed estimator outperforms the adopted estimators. Overall, we discovered that the suggested estimator performs better than all of the already used estimators.