Statistical estimates bias for data with missing values: Nonresponse study

انحياز التقديرات الإحصائية للبيانات التي تحوي قيماً مفقودة: دراسة حالة عدم الاستجابة

الملخص

Statistical surveys are usually conducted to obtain data describing a problem in a studied society, and many surveys experience a rise in nonresponse rates, as the rate of nonresponse may affect the bias of the nonresponse in survey estimates. Recent empirical results show instances of nonresponse rate correlation with nonresponse bias, we attempt to translate statistical experiences of nonresponse bias in newly published studies and research into causal models that lead to assumptions about when a lack of response causes bias in estimates.

Research studies of the estimates of nonresponse bias show that this bias often exists. The logical question is: what is the advantage of surveys if they suffer from high rates of nonresponse, since post-survey adjustments for nonresponse require additional variables, the answer depends on the nature of the design and the quality of the additional variables.