STUDYING THE EFFECT OF NON-RESPONSE ON IMPROVED ESTIMATORS AND THEIR EFFICIENCY IN THE PRESENCE OF SIMULTANEOUS NON-RESPONSE AND MEASUREMENT ERRORS

In survey sampling, it is a common practice that two non-sampling errors do occur during collection of survey data. These errors are the non-response and measurement errors. Non-response error is the most serious error in survey sampling. Estimation strategies such as Ratio, Product and Regression methods are influenced by these errors. Also, the properties of these estimators are influenced significantly. So, obtaining estimators that are less affected by these errors is a challenging task. It is the responsibility of the researcher(s) to develop estimators that are less affected by these errors to obtain more valid, reliable, and efficient estimators. In this paper, existing improved mean estimators studied under the influence of measurement errors alone will be restudied under the effect of non-response alone and their efficiency in the presence of non-response and measurement errors. The properties (Biases and mean square errors) were derived up to first order approximation using Taylor series method. A numerical study was conducted to ascertain the performance of the proposed estimators under the effect of non-response and measurement errors and the result revealed that the proposed estimators are efficient. Keywords: Estimate, Estimation, Measurement Errors, Non-response, Survey