Abstract: In this paper a two step calibrated estimator is proposed for reducing both sampling error and nonresponse bias in a survey. The two steps are motivated by the typology of available auxiliary information. The information related to the study variable is separated by that related to the response process. At first the sampling weights are modified in order to respect the sampling estimates of the totals of the variables linked to the response process. The intermediated weights are again adjusted in order to respect the population totals of both the typology of auxiliary information. The mean squared error of the estimator is studied separating the contribution of the variance from that of the bias. A simulation study provides evidence about the empirical properties of the proposed estimator in comparison with the traditional calibrated estimator.
Stimatori calibrati in presenza di non risposta / M. PRATESI; E. ROCCO. - STAMPA. - (2006), pp. 329-339.
Stimatori calibrati in presenza di non risposta
ROCCO, EMILIA
2006
Abstract
Abstract: In this paper a two step calibrated estimator is proposed for reducing both sampling error and nonresponse bias in a survey. The two steps are motivated by the typology of available auxiliary information. The information related to the study variable is separated by that related to the response process. At first the sampling weights are modified in order to respect the sampling estimates of the totals of the variables linked to the response process. The intermediated weights are again adjusted in order to respect the population totals of both the typology of auxiliary information. The mean squared error of the estimator is studied separating the contribution of the variance from that of the bias. A simulation study provides evidence about the empirical properties of the proposed estimator in comparison with the traditional calibrated estimator.I documenti in FLORE sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.