Resumo

CUESTA, Cristina; MARI, Gonzalo  e  RIVERO, Matias. Estimación de la función generalizada de variancias para la encuesta anual de hogares de la ciudad autónoma de Buenos Aires, año 2010. SaberEs [online]. 2013, vol.5, n.2, pp.00-00. ISSN 1852-4418.

Generalized Variance Functions (GVF) is a highly disseminated methodology among government agencies entrusted with planning and implementing complex surveys. This method allows obtaining the sample error of estimation in a simple way. It is a model which main objective is to reproduce the relationship between an estimator and a measure of its variance, for example, the Relative Variance (RV). It is possible to adjust several models for de GVF but in practice two simple models are usually considered (linear and logarithmic models). In this article we explore the use of other model based in P-spline regression. Models are compared using data from the General Direction of Statistics and Census of Buenos Aires City. The estimations of totals of characteristics from "Annual Home Survey" of 2010 will be used.

Palavras-chave : Estimation errors; Complex survey designs; P-spline models.

        · resumo em Espanhol     · texto em Espanhol     · Espanhol ( pdf )