瑞士人口普查覆盖调查的抽样设计与启示

Sampling Design of the Swiss Census Coverage Survey and Its Enligh enment

瑞士2000年首次独立进行人口普查覆盖调查。覆盖调查由抽取建筑物样本、计数建筑物样本住宅、抽取建筑物子样本以减少家庭规模、调查样本家庭和估计人口普查净遗漏率、遗漏率和错误计数率构成。覆盖调查抽样设计分北部和南部分别进行。抽样分两步:第一步抽取建筑物初始样本,第二步从抽取的初始建筑物样本中抽取家庭。北部初始建筑物样本通过三个阶段完成,采取分层不放回不等概率方式抽取。南部初始建筑物样本通过两个阶段完成,采取与北部同样的方式抽取。瑞士方案对我国具有启迪和借鉴作用。

Switzerland carried out its first independent census coverage survey in 2000. Coverage survey consists of drawing the sample of buildings, counting dwellings in the sample of buildings, drawing the sub-sample of buildings to reduce the family size, interviewing sample families, and estimating census net omission rate, omission rate, and erroneous enumeration rate. Sampling design of the coverage survey was conducted respectively in the north and the South. Sampling included two steps： the first step was to draw initial sample of buildings, while the second was to draw families samples from the initial samples of buildings. Sampling for initial samples of buildings in the north was completed in three phases by means of adopting stratification without replacement and unequal probability. Sampling for initial samples of buildings in the south, within two stages, was completed in the same way as those applied in the north. The project discussed in the article is thus considered to provide enlightenment and reference to that of our country.