414 – Small-Area Estimation: Applications and Enhanced Methods
Small-Area Estimation in Household Surveys When Auxiliary Variable Totals Are Known
Prabhakar Ghangurde
Consultant
Assuming standard regression model for one small area domain (and its complementary domain) with known auxiliary variable population totals BLUE of domain total of an estimation variable is derived as regression-synthetic estimator with known auxiliary variable population totals instead of domain means in the estimator based on type-B model. Assuming one auxiliary variable in the model BLUE of domain total of an estimation variable is derived. It reduces to synthetic estimator with ratio-adjustment (Ghangurde & Singh (Survey Methodology(1977)) and then to ratio-synthetic estimator; the latter reduction involves a substitution of synthetic weights in terms of administrative data. It is assumed that the ratio of estimation variable total to auxiliary variable total in the domain is equal to the ratio in the population. The assumption is made explicitly in synthetic estimator but is implicit in ratio-synthetic estimator.
