by Stefan, Marius
Published in Romanian Journal of Economic Forecasting, 2008, volume 9 issue 2, 67-91.
| Requires a PDF viewer such as Xpdf
In our paper we show how to construct a model for one variable in the French Health Survey data set: the number of times an individual visited a generalist in the last twelve months, for which we are interested in estimating the regional means. Then, we test the fit of the model to the data and compare it to other two alternative models. We derive theoretical formulas for the estimates of the twenty-two regional means along with their standard deviations. We compare this to the design-based estimations obtained by INSEE in the case of the five regions with extra sample. We discuss some alternative for future research.
small areas, direct and indirect estimations, Markov chains, Gibbs sampling, Metropolis-Hastings algorithm
JEL Classification: C12, C13, C46