【摘要】：文章從分層抽樣的角度,以設(shè)計(jì)效應(yīng)為工具,構(gòu)造出若干種分層抽樣的設(shè)計(jì)效應(yīng)模型,應(yīng)用于分層抽樣的樣本分配設(shè)計(jì)。由于分層設(shè)計(jì)效應(yīng)主要取決于各層平均數(shù)的差異,則差異越大,分層的效果越好。為了滿(mǎn)足目標(biāo)變量全國(guó)估計(jì)量和特定小域估計(jì)量的精度要求,樣本在各小域的最優(yōu)分配可能是比例分配和非比例分配的折中。
[Abstract]:From the point of view of stratified sampling and with the tool of design effect, this paper constructs several design effect models of stratified sampling, which are applied to the sample allocation design of stratified sampling. Because the effect of layered design mainly depends on the difference of the average of each layer, the greater the difference, the better the effect of stratification. In order to meet the accuracy requirements of the national Estimator of the target variable and the specific subdomain Estimator, the optimal distribution of the samples in each subdomain may be a compromise between the proportional distribution and the non-proportional distribution.
【作者單位】： 廣東工業(yè)大學(xué)管理學(xué)院;廣東工業(yè)大學(xué)大數(shù)據(jù)戰(zhàn)略研究院;
【基金】：國(guó)家社會(huì)科學(xué)基金資助項(xiàng)目(17BTJ037) 廣東省哲學(xué)社會(huì)科學(xué)“十二五”規(guī)劃學(xué)科共建項(xiàng)目(GD15XGL09)
【分類(lèi)號(hào)】：C31
，

本文編號(hào)：2514502