Nonparametric regression method for estimating the error variance in unistage sampling

Abstract

Nonparametric regression provides computationally intensive estimation of unknown finite population quantities. Such estimation can be more robust than inference tied to model based inference. A nonparametric procedure for estimating error variance is suggested. An empirical example is given to illustrate the performance of the derived estimator vis-a-vis the currently popular variance estimator in model based surveys.