Evaluation of Efficiency of Standard Regression Mixed Model--Based BLUP Estimators in Household Surveys Assuming Unequal Error Variances

• Activity Number: 467
• Type: Contributed
• Date/Time: Wednesday, August 6, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Survey Research Methods Section
• Author(s): Prabhakar Ghangurde*+
• Keywords: M.S.E.Efficiency ; Mixed ; Model-based ; BLUP ; Unequal ; ErrorVariances

Abstract

Assuming standard regression mixed model with unequal error variances and cluster sampling of households m.s.e. efficiency of BLUP estimator,as compared to ratio-synthetic estimator, is evaluated. A sample of given number of households is selected in two stages,clusters and households, from a population of geographic strata with two domains: domain of interest Ui and complementary domain Uc (see Ghangurde P.D.; JSM (2013)). Sample size in Ui is increased in steps of number of households and that in Uc is decreased in steps of the same number of households. In a household survey BLUP estimation was done using a quantitative auxiliary variable assumed > 0 (e.g. annual household income ). Efficiencies were obtained for several harmonic means, total sample sizes and sample sizes in Ui and Uc. The results show that ratio of harmonic mean to sample size in Ui and ratio of total sample size to that in Ui essentially determine efficiency of BLUP as compared to ratio-synthetic estimator; ratio of harmonic mean to sample size in Uc is not important .Data on efficiencies obtained by using assumed values of harmonic means and sample sizes in Ui and Uc are presented and analyzed.