USE OF ESTIMATING FUNCTIONS FOR ESTIMATION FROM COMPLEX SURVEYS

We describe a method for point and interval estimation of population p arameters from complex surveys using estimating functions. The theory was originally developed for infinite populations and has recently bee n applied to finite populations. With estimating functions, a unifying framework can be given for point and interval estimation of both fini te and infinite population parameters. We discuss test inversion metho ds to derive confidence intervals for one-dimensional parameters and p ropose a method for eliminating nuisance parameters in the multidimens ional setting. We show that special cases of our proposal result in co nditional and orthogonal methods proposed in the literature. We descri be a simulation study using real data to compare the coverage probabil ities of confidence intervals obtained under various approaches.