论文部分内容阅读
在对抽样调查数据进行分析时经常遇到无回答的情况,无回答问题已经成为影响数据质量的一个重要因素。现实中导致无回答的原因很多,主要来自于调查中的无回答、遗漏调查项、不合理数据剔除等因素。无回答的存在会引起调查估计的偏倚,如果无回答率很高,偏倚将会严重到使调查失败,而且这种偏倚不会随着样本量的增加而减小,抽样方差也会增大(鲁志贤,2002)。所以处理好无回答问题,对于提高数据的质量和对数
In the analysis of the sample survey data often encountered no answer, no answer has become an important factor affecting data quality. There are many reasons that lead to no answer in reality, mainly due to the non-response in the survey, omitted survey items, unreasonable data exclusion and other factors. The existence of an unresponsiveness can cause bias in survey estimates. If the non-response rate is high, the bias will be so severe that the survey will fail, and the bias will not decrease with the sample size and the sample variance will increase ( Lu Zhixian, 2002). So deal with non-answer questions, to improve the quality and logarithm of the data