Assessing Auxiliary Vectors for Control of Nonresponse Bias in the Calibration Estimator
Assessing Auxiliary Vectors for Control of Nonresponse Bias in the Calibration Estimator
复制标题
评估校准估计器中控制无响应偏差的辅助向量
DOI：
--
复制
发表时间：
2008
期刊：
影响因子：
--
通讯作者：
Sixten Lundström
中科院分区：
文献类型：
--
作者：
C. Särndal;Sixten Lundström
Statistics Sweden 7 Abstract This paper deals with calibration estimation for surveys with nonresponse. Efficient weighting adjustment for unit nonresponse requires powerful auxiliary information. The theory in the paper is inspired by the survey environment in Scandinavia (and in other North European countries), where many reliable administrative registers provide rich sources of auxiliary variables, in particular for surveys on individuals and households. The weights in the calibration estimator are computed on information about a specified auxiliary vector. Even with the “best possible” auxiliary vector, some bias remains in the estimator. A close approximation to the remaining bias is presented and analyzed. The relationship between the bias expression and the auxiliary vector in use is a focal point in the article. The many potential auxiliary variables allow the statistician to compose a wide variety of possible auxiliary vectors. The need arises to compare these vectors to assess their effectiveness for bias reduction. To this end we define and examine an indicator useful for ranking alternative auxiliary vectors in regard to their ability to reduce the bias. The indicator is computed on the auxiliary vector values for the sampled units, responding and nonresponding. An advantage is its independence of the study variables, of which there are many in a large survey. The properties of the indicator are examined in the theory sections of the paper. The indicator tends, with increasing sample size, to a population analogue, shown to be linked to the bias through an approximately linear relationship. The higher the value of indicator, the more likely it is that the bias will be low, for many study variables. Empirical studies occupy the final sections of the paper. A synthetic population is constructed and potential auxiliary vectors are ranked with the aid of the indicator. Another empirical illustration illustrates how the indicator is used for selecting auxiliary variables in a large survey at Statistics Sweden. Assessing Auxiliary Vectors for Control of Nonresponse Bias in the Calibration Estimator 8 Statistics Sweden
