GWSM and Calibration
The GWSM described in the previous chapters does not use auxiliary information to obtain estimation weights. It can be imagined, however, that the use of auxiliary variables can improve the precision of estimates coming from the GWSM. For example, auxiliary information could come from the population U A from which the sample is selected, from the target population U B , or both of the populations. In this chapter, we are going to show that it is possible to associate the calibration of Deville and Särndal (1992) to the GWSM. In fact, it corrects the estimation weights from the GWSM so that the estimates produced correspond to known totals associated to auxiliary information. We will show that it is possible in this case to use auxiliary information from the two populations U A and U B . We will develop in particular the regression estimator coming from the application of calibration.
KeywordsDistance Function Auxiliary Variable Asymptotic Variance Auxiliary Information Asymptotic Bias
Unable to display preview. Download preview PDF.