The Resampling Method for Estimating Variance of the Generalized Regression Estimator in the Presence of Nonresponse

บทคัดย่อ

This paper aims to propose new variance estimators for the generalized regression estimator to estimate the population mean under a reverse framework, where nonresponse occurs in the study variable, as proposed by Ponkaew [1] in 2018. The proposed variance estimators are investigated using two resampling techniques, namely the Jackknife and the Rao-Wu bootstrap. The new variance estimator does not require joint inclusion probabilities, which differs from the variance estimator proposed by Ponkaew [1]. The effectiveness of the suggested estimators is examined through simulation experiments and an application to air pollution data from Phetchabun province, Thailand. The findings demonstrate that, in comparison to other estimators, the proposed Rao-Wu bootstrap variance estimator achieves the highest precision, producing the smallest the root mean square error.