在輪換抽樣設計的調查裡，可以利用樣本由某一時期到下一個時期的時間重疊來降低對前後時期改變量估計的變異數。為了顯現重複抽樣的優點，複合估計量被用來結合當月以及之前那些月份的資料。在這篇文章中，我們提出一個計算廣義遞迴性複合估計量之最佳權數組的程序。同時，討論一個運用有限移動運算子所建構的更一般化複合估計量，並且以美國人口普查局的人口現況調查資料在美國各州勞動力的估計為例來和遞迴性複合估計量相比較。

In a survey with rotating design, the overlap from one time period to the next can Teduce the variance of estimators of month-to-month change. In order to take advantage of repeated sampling, composite estimators are used to combine data for the current month with those from prior months. In this paper, we provide procedures for computing a set of optimal weights for the generalized recursive composite estimator. A more general composite estimator constructed with a finite moving operator is defined and compared to the recursive compositor in the estimation of the state labor force from the CPS data set.