中文摘要：

在人口普查质量评估中,为了计算人口普查误差,通常先要用双系统估计量来构造一个实际人口数的估计量。而双系统估计量要求在被普查登记概率相同的同质人口总体中构造。为此,需要对人口普查质量评估中抽取的样本中的人按照影响人口被登记概率的诸种变量进行抽样后交叉分层,本文讨论了应用优比选择抽样后分层变量的理论与实践问题。

英文摘要：

It is usually necessary to construct an actual population estimator by means of dual system estimator in order to calculate the census error in the census quality assessment firstly.The dual system estimator must be constructed in a homogeneous population whose people own the same registered probability in the census.To this end,people of the sample drawn from census quality assessment should be carried out crossing stratification after sample is drawn according to some variables that affect population registration probability.This paper discusses some problems of theory and practice on using dominance ratio to choose stratification variables after sampling.