序贯抽样在计算机辅助电话调查中的应用

样本量的确定是抽样设计中的关键问题,传统方法利用总体方差和调查费用的有关信息来确定样本量可能产生两种结果,一种是样本量过低,无法保证希望的估计精度要求;一种是样本量过高,导致调查经费的浪费。计算机辅助电话调查中即时的数据运算和管理功能为序贯抽样的应用奠定了基础。利用前期抽取样本的计算结果,可以规定进一步需抽取的样本量,最终样本量是对真正期望样本量的一个最佳近似,它比传统方法更能保证以最少的费用满足预先设定的精度要求。

Application of Sequential Sampling in Computer-Assisted Telephone Interview

Determining the sample size is crucial for a survey sampling design,for which traditional approaches have to seek some information on population variance or survey costs,however,which may result in either sample size too small and real precision undre desired level,or sample size too large and waste of survey costs.Computer-assisted telephone interview does make it possible to apply sequential sampling in large social survey because of its function of instant data operation and management.Using calculation result from prophase sample,sampling size needed in next step can been determined and it can be expected that ultimate sample size would be a best approximation for that desired and meet the precision demanded at less cost than traditional methods.