A New Modified Estimator for Population Mean using Auxiliary Character and Attribute in Double Sampling

In sample surveys, when information on two auxiliary characters is available, several authors prefer chain-type estimators to improve the performance of estimators of the population mean. These authors have assumed that the information (population mean, total, etc.) on the first auxiliary or auxiliary character is not available but on the second or additional auxiliary character. However, there are instances where the latter and former may also be unavailable. In such cases, we have suggested a novel modified regression-type estimator that uses the data from the auxiliary character and the auxiliary attribute in double sampling to estimate the population mean. We have derived the expressions for bias and mean square error of the suggested estimator using the large sample approximation. After mathematical comparisons, we have obtained mathematical conditions under which the suggested estimator has lower mean square error than the existing estimators. Further, we supported these theoretical comparisons using a numerical analysis with a real-life dataset.. KEYWORDS :Bias, Double sampling, Mean square error, Auxiliary attribute, Auxiliary character.