【摘要】 人口普查不可避免产生误差,包括涵盖误差和内容误差。涵盖误差可进一步分为净涵盖误差及遗漏和错误计数。一般是先估计净涵盖误差和错误计数,然后估计遗漏。净涵盖误差为真实人口数与普查人口数之差。目前世界上大多数国家是使用人口普查后的事后计数调查来估计真实人口数。事后计数调包括抽样设计、调查样本、比较样本、事后分层和基于捕获-再捕获模型的双系统估计量的构造。我国每次人口普查后都进行了事后计数调查和估计了净涵盖误差,但我国的做法与其他国家相比存在某些缺陷,需要改进与完善。本文对我国事后计数调查提出若干对策建议。更多还原

【Abstract】 Population census contains inevitable errors,including coverage error and content error.Coverage error is further classified into net coverage error,erroneous enumerations and omissions.Generally,net coverage error and erroneous enumerations are first estimated and then omissions are estimated.Net coverage error is difference between true and census population count.At present most countries use the post-enumeration survey after census to estimate true population count.The post-enumeration survey includes sample design,sample survey and comparison,post-stratification,and dual system estimator construction based on capture-recapture model.China carried out post-enumeration survey after each census and estimate net coverage error.However compared with methods used by other countries the Chinese one is beset with some drawbacks and need improvement.This paper puts forwards some advices on post-enumeration survey for China.更多还原