GWR-assisted integrated estimator of finite population total under two-phase sampling: a model-assisted approach

AbstractIn survey sampling, auxiliary information is used to precisely estimate the finite population parameters. There are several approaches available in the literature that provide a practical method for incorporating auxiliary information during the estimation stage. In order to effectively utilize the auxiliary information, a geographically weighted regression (GWR) model-assisted integrated estimator of finite population total under a two-phase sampling design has been proposed in this article. Spatial simulation studies have been conducted to empirically assess the statistical properties of the proposed estimator. In the presence of spatial non-stationarity, empirical findings reveal that the proposed estimator outperforms all existing estimators such as two-phase HT, ratio, and regression estimators, demonstrating the importance of spatial information in survey sampling.KEYWORDS: Data integrationgeographically weighted regressionmodel-assisted approachspatial non-stationaritytwo-phase regression AcknowledgementThe authors are thankful to the blind reviewers for providing valuable suggestions that have greatly enhanced the quality of the article. The first author would like to express his heartfelt gratitude to the ICAR-Indian Agricultural Statistics Research Institute, New Delhi, India, for providing the real CCE survey data, lab facilities, and overall support to conduct the research work during Ph.D. programme.Disclosure statementNo potential conflict of interest was reported by the authors.Data availability statementData sharing is not applicable.