A note on estimating the variance of the regression estimator

Abstract

SUMMARY Särndal, Swensson & Wretman (1989) and Kott (1990a) have proposed two separate estimators for the variance of the regression estimator of a finite population mean based on a simple random sample without replacement. It is shown here that the former approximates for large samples and the latter equals respectively two known variance estimators, relative merits of which are already studied in the literature.