Model-based variance estimation under unequal probability sampling

Abstract

Horvitz–Thompson estimator, Model-based estimation, Variance estimation,

A Monte Carlo comparison of some variance estimators of the Horvitz–Thompson estimator

“…This agrees with the results of Patel and Chaudhari[12]. (3) The estimators V CYG , V RYG , and V IYG based on V YGY turned out to be consistently more efficient than the estimators based on V HT Y .…”

“…large populations 6-10 and z and x have very high correlation with y.) (7) V CHT has performed better than all estimators when x and y variables are perfectly correlated (see populations 11,12). Also, its absolute RBs are within 1% for these populations.…”

“…This has prompted us to investigate the same problem using model-based and design-based approaches. Some recent references on HT variance estimation include Stehman and Overton [10], Berger [11], and Patel and Chaudhari [12] among others.…”