Options
Variance estimation in the presence of imputed data for high entropy sampling designs
Auteur(s)
Haziza, David
Maison d'édition
Neuchâtel Université de Neuchâtel Institut de Statistique
Date de parution
2014
Nombre de page
22
Résumé
Variance estimation is an important aspect of the estimation process in statistical agencies as variance estimates provide a measure of the quality of the survey’s estimates. In the presence of imputed data, usual variance estimators rely on the availability of the second-order inclusion
probabilities, which may be difficult (even impossible) to compute for ome sampling designs. In this paper, we derive simplified variance estimators that result from approximating the second-order inclusion probabilities in terms of the first-order inclusion probabilities. Results
of a simulation study, evaluating the properties of the proposed variance estimators in terms of bias and mean squared error, will be presented.
probabilities, which may be difficult (even impossible) to compute for ome sampling designs. In this paper, we derive simplified variance estimators that result from approximating the second-order inclusion probabilities in terms of the first-order inclusion probabilities. Results
of a simulation study, evaluating the properties of the proposed variance estimators in terms of bias and mean squared error, will be presented.
Type de publication
Resource Types::text::working paper