Linear mixed models underpin many small area estimation (SAE) methods. In this paper we investigate SAE based on linear models with spatially correlated small area effects where the neighbourhood structure is described by a contiguity matrix. Such models allow efficient use of spatial auxiliary information in SAE. In particular, we use simulation studies to compare the performances of model- based direct estimation (MBDE) and empirical best linear unbiased prediction (EBLUP) under such models. These simulations are based on theoretically generated populations as well as data obtained from two real populations (the ISTAT farm structure survey in Tuscany and the US Environmental Monitoring and Assessment Program survey). Our empirical results show only marginal gains when spatial dependence between areas is incorporated into the SAE model.
Small area estimation for spatially correlated populations. A comparison of direct and indirect model-based methods
SALVATI, NICOLA;
2007-01-01
Abstract
Linear mixed models underpin many small area estimation (SAE) methods. In this paper we investigate SAE based on linear models with spatially correlated small area effects where the neighbourhood structure is described by a contiguity matrix. Such models allow efficient use of spatial auxiliary information in SAE. In particular, we use simulation studies to compare the performances of model- based direct estimation (MBDE) and empirical best linear unbiased prediction (EBLUP) under such models. These simulations are based on theoretically generated populations as well as data obtained from two real populations (the ISTAT farm structure survey in Tuscany and the US Environmental Monitoring and Assessment Program survey). Our empirical results show only marginal gains when spatial dependence between areas is incorporated into the SAE model.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
