On asymptotically design-unbiased estimators of a finite population mean
SUMMARY: Certain regression-type estimators of a finite population mean are examined for their asymptotic design properties. A sufficient condition that implies the design-consistency or asymptotic design-unbiasedness is discussed for a general regression estimator. The leading term of its asymptotic bias is explicitly derived when this condition holds as well as when it does not hold. © 1991 Biometrika Trust.
Deng, L., & Chhikara, R. (1991). On asymptotically design-unbiased estimators of a finite population mean. Biometrika, 78 (1), 189-195. https://doi.org/10.1093/biomet/78.1.189