Abstract
Sampling variances are compared for the conventional and for two classes of multiplicity estimators based on stratified sample surveys. The conventional estimator uniquely links every individual to a single enumeration source—its index source. The multiplicity estimators link every individual to its index source, and possibly to other sources as well. One multiplicity estimator permits an individual to be linked to sources in one stratum only; the other multiplicity estimator permits an individual to be linked to any sources. None of the three estimators is necessarily more reliable than the others.

This publication has 0 references indexed in Scilit: