敏感性问题的抽样调查中非随机化响应技术的新进展

Advances of the Non-Randomized Response Techniques in Sample Surveys with Sensitive Questions

在社会学、经济学、医学、心理学或政治学的许多抽样调查中,往往会涉及到一些具有敏感性的调查项目.如果直接询问受访者,由于涉及到个人隐私而使得他们拒绝回答甚至报告错误的答案.因此,发展合适的统计模型以鼓励受访者提供真实的回答且保护他们的隐私显得尤为重要.非随机化响应技术相较于传统的随机化响应技术具有可重复性、节省成本、应用范围广等优点.本文对近十年来提出的九种非随机化响应技术进行了回顾和比较,可为研究人员进行敏感性信息搜集与分析提供有用的选择.

In many social, economic, medical, psychological and political surveys, they often involve some highly sensitive questions. Since such questions refer to individuals’ privacy and may make them feel embarrassed, they are likely to refuse to respond or, even worse, provide false answers if been asked directly. Therefore, it is of great important to develop appropriate models to encourage truthful answers as well as protecting privacy. Compared to the conventional randomized response techniques, the nonrandomized response approaches have advantages such as the reproducibility, the cost economized, wide range of application and so on. In this paper, we make a brief review and comparison for the nine newly developed non-randomized response approaches during the last decade, aiming at guiding researchers to choose the most appropriate ways to collect sensitive information.