摘要
人口统计分析模型和双系统估计量是美国和其他一些国家目前提供普查净误差估计值的两个基本方法,未来可能使用三系统估计量做这项工作。采用文献解读法、拓展研究法、调查研究法和抽样估计法综述这三个估计量形成的历史背景、理论依据及模型表达式。研究结果表明：人口统计分析模型适合于提供年轻孩子组的实际人数估计值,但行政记录系统与人口普查对种族分类的不一致影响其估计的精度及与其他来源数据的可比性;双系统估计量要求两个系统独立,否则估计的净误差内含系统性偏差;三系统估计量不要求系统之间独立,同时利用了人口普查及人口行政记录辅助信息,将在未来人口普查净误差估计中发挥主导作用。
The demographic analysis model and the dual system estimator are the two basic methods available in the United States and some other countries, which provide an estimate of the net error of the census. Triple system estimator may be used in the future. This paper reviews the historical background, theoretical basis and model expression of the above three estimators by means of literature interpretation, expanded research, investigation and sampling estimation. The results show that： The demographic analysis model is suitable for providing the actual number estimates of young children group,but the racial classification difference between the system of administrative records and census affects the estimation precision and comparability with other sources of data; dual system estimator requires two systems independent, otherwise net error estimated includes systematic bias; triple system estimator will play a leading role in the future estimation of the census net errors because it does not demand three independent systems, and treats administrative records and population census as supplementary information.
作者
胡桂华
HU Gui-hua(School of Math and Statistics,Chongqing Technology and Business University,Chongqing 400067 China;Population Centre,Minnesota University,Minneapolis,MN55455,The United States)
出处
《数理统计与管理》
CSSCI
北大核心
2018年第5期796-814,共19页
Journal of Applied Statistics and Management
基金
2016年重庆市社会科学规划重大委托项目(2016WT03)
国家社会科学基金项目(15BTJ011)
全国统计科学研究重点项目(2016LZ29)
2017年重庆高校人文社会科学研究重点项目(17SKG088)
国家自然科学基金面上项目(11471239)
美国儿童健康和人类发展研究院、美国健康研究院及人口统计和行为科学项目(R01HD047283)
美国自然科学基金项目(SES-0851414)
