全面调查,还是抽样调查?——基于调查效率比较的方案设计决策

Complete Survey or Sample Survey?——Survey Design Decision Based on Efficiency Comparison

文章构建了基于均方误差分解的全面调查与抽样调查之间效率比较的理论分析框架。通过设定两种调查方式下调查精度的一般化公式,以及调查误差与调查成本的组合函数,藉此分析确定对两种调查方式进行取舍的成本-误差临界点,并以简单总体为例对效率比较的过程进行了演示。分析表明,如果全面调查的误差规模较大(调查精度较差),而抽样调查中用于改进调查质量的单位成本增量较小时,可以考虑以抽样调查代替全面调查,以求获得更高的调查效率。

In order to compare the efficiency of complete survey and sample survey,a theoretical analysis framework based on mean square error（MSE） decomposition has been constructed.By deducing the survey precision expressions of the two survey methods and designing the cost-error function,the cost-error critical point for survey methods selection can be ascertained,and a demonstration with simple population has been given.The paper argues that the sample survey would have higher survey efficiency than the complete survey,under the condition that complete survey suffers severe error disturbance and sample survey has small unit cost increment for improving survey quality.