摘要
针对长问卷存在无回答率和回答负担,从而导致统计调查精度降低的问题,采用问卷分割法解决该问题,并且通过小域估计的方法进行参数估计。模拟研究表明,利用小域估计方法对分割问卷进行参数估计显然优于用多重插补法进行参数估计。研究结果表明,运用小域估计方法对分割问卷进行参数估计,能显著提高统计调查的精度。
Because of the problem of lengthy survey questionnaires have non-response question and response burden ,leading to reduce statistical survey precision .Then we use split questionnaire design to solve this problem ,and through small area estimation for parameter estimation .In a simulation study ,we show that small area estimation provides more reliable statistics than multiple imputation methods .Finally draw the conclusion ：using small area estimation method for split questionnaire’s parameter estimation can significantly increase the estimation accuracy .
出处
《统计与信息论坛》
CSSCI
2014年第10期14-18,共5页
Journal of Statistics and Information
关键词
回答负担
矩阵抽样
经验最佳线性无偏预测
多重插补
response burden
matrix sampling
empirical best linear unbiased prediction
multiple imputation
