The Implementation of Experimental Surveys for Data-Quality Improvement.

为了抑制非反应和反应错误，组织了两次实验调查。我们假设了影响反应行为的两个因素：成本感知和风险感知。遗憾的是，由于实验条件设置不充分，成本知觉测试失败。风险感知测试显示，人们倾向于在他们认为低风险的情况下做出反应。然而，这种差异相当微小，不足以检验我们的假设，即风险感知会因人口统计学变量而变化，从而导致无反应偏差的出现。

Aiming at repressing non-response and response errors, two experimental surveys were organized. We assumed two factors that affect response behaviors : cost-perception and risk-perception. Unfortunately, the cost-perception test failed due largely to inadequate experimental condition setting. The risk-perception test revealed that people tend to respond as they perceive low risk to answer. However, the difference was rather tiny and inadequate to examine our hypothesis that the risk-perception would vary by demographic variables which results in the appearance of non-response bias.