Informative Sampling on Two Occasions: Estimation and Prediction

Abstract

The sample distribution is defined as the distribution of the sample measurements given the selected sample. Under informative sampling, this distribution may be different from the corresponding population distribution. Sampling on two occasions under informative sampling design, utilizing the sample and sample-complement distributions for occasion one, the matched sample and unmatched sample distributions, and matched sample-complement and unmatched sample-complement for occasion two, is proposed for predicting finite population total of a variable under study for the current (second) occasion, viewing information collected on the first (previous) occasion as auxiliary information. An interesting result of the present analysis is that known predictors in common use are shown to be special cases of the present predictors obtained under informative sampling, thus providing them a new justification.