Comparación de estimadores directos e indirectos en estimación de áreas pequeñas

Abstract

Estimating in small areas allows obtaining more accurate and reliable results in domains whose simple sizes are zero or close to zero. Di_erent estimators that depend on the sampling design as the direct estimators, estimators proposed by the theory to obtain better results in domains where the sample size is small, such as synthetic estimators and estimators using mixed linear models for small areas, are the topics which are presented in this degree work, where the behavior of each one of them is observed, to be compared through the mean square error and in this way to obtain conclusions about their performance when they are used in those areas where the sample size is small.