论列举实验在敏感问题调查中的应用——以非制度化政治参与为验证

List Experiment in Measuring Sensitive Issue——Example from Non-institutionalized Political Participation

敏感问题测量是社会科学实证研究的难点,基于社会意愿偏差,人们对于敏感问题或者选择不应答,或者选择非真实应答。随着调查研究技术的进步,近年来,社会科学界为此掀起讨论热潮,列举实验是人们讨论的热门方法之一。列举实验的关键在于避免产生设计效应、天花板效应和地板效应。本文以非制度化政治参与的调查数据分析为基础,验证显示列举实验可以降低无应答比例,提高真实应答比例,并降低社会意愿偏差。而将列举实验分别作为自变量和因变量的示例分析表明,相比直接提问方式,列举实验对模型估计具有关键影响,它可以提供较为可靠、可信和恰当的分析结果。当然,列举实验仍然面临敏感问题测量的本质困难、实验效力、设计和分析难点等问题,因此需要进一步优化改进。

It is difficult to measure sensitive issues in social science study, for respondents choose not to report or lie to the questions as a result of social desirable bias. With the development of survey technology, such a difficulty comes back to academic discussion in recent years, in which list experiment is three most popular methods. The key of list experiment design is to avoid design effect, ceiling effect and floor effect. This study takes non-institutionalized political participation as an example. We introduce the method to test the assumption. And we prove that list experiment can lower the nonresponse rate, raise the truthful response rate and lower the social desirable bias. Compared with direct question, list experiment has significant impacts on the model estimation, which provide more reliable, credible and appropriate result, when applying list experiment as dependent and independent. Of course, list experiment still has to face the problems include the essential problem of sensitive issues measuring, power, design and analysis, which means it still need further improvement.