Predicting the Finite Population Distribution Function under a Multilevel Model

Chambers and Dunstan proposed a model-based predictor of the population distribution function that makes use auxiliary information under general sampling design. Subsequently, Rao, Kovar, Mantel design-based ratio difference predictors also this information. Both (CD RKM) assume single level model for target population. In article we develop finite follows multilevel model. These new same smearing approach underpinning CD predictor. We compare our with RKM via simulation, show they perform better than these when there is significant intra-cluster correlation. The performances two are examined an empirical study based on data from large-scale UK business survey aimed at estimating hourly pay rates. AMS subject classification: Primary 62G30, Secondary 62G32