论双系统估计量的无偏性

Discussing Unbiasedness of Dual System Estimator

人口普查质量评估中所使用的双系统估计量是否为无偏估计量是一个很值得深入讨论的问题。只有无偏,才能确保使用双系统估计量估计的目标总体实际人数及人口普查净误差平均等于它们的实际数。针对人口普查质量评估工作中所使用的双系统估计量,论证这个估计量的无偏性条件。采用从既定假设出发进行推演的路径论证。研究结果表明,双系统估计量是目标总体实际人数无偏估计量的必要但非充分的条件是,人口普查与其质量评估调查相互独立以及目标总体中的每一个人在人口普查中的登记概率相同,在质量评估调查中登记的概率也相同。

Whether dual system estimator used in census quality assessment is an unbiased estimator is a very worthy of further discussion.Only unbiased,can we ensure population true count and census net error estimated by dual system estimator equal to their true count on average.Against the dual system estimator used in census quality assessment,proving unbiased conditions of the estimator.The path from a given hypothesis to deduction is used to prove.The results show that necessary but not sufficient condition that dual system estimator becomes unbiased estimator of target universe true population is independent between census and its quality assessment survey,and everyone in the universe has the same probability of being registered in the census or quality assessment survey.