ESTIMATING THE DIFFERENCE OF MEANS WITH IMPUTATION OF THE MISSING OBSERVATIONS

We consider the use of a finite population model for estimating the difference of the means of two variables under amechanism for missing observations. The information available on the two variables is used for imputing thecorresponding missing data. The proposed estimator is characterized by means of the expected squared error. Data from anenvironmental study are used for illustrating the performance of the estimator.