Estimation and Seasonal Adjustment of Population Means Using Data From Repeated Surveys

Abstract

I consider estimation and seasonal adjustment of population means based on rotating panel surveys carried out at regular time intervals. The analysis uses a dynamic structural model that assumes a decomposition of the mean into a trend-level component and a seasonal component. The model accounts for the correlations between individual panel estimators and for possible rotation group effects. It can be applied to general rotation patterns using either the individual panel estimates or the aggregate sample estimates, depending on data availability. Empirical results illustrating the important features of the procedure are presented.