实证社会科学研究中的因果推断:挑战与精进  

Causal Inference in Empirical Social Science Research:Challenges and Refinements

作　　者：王正绪 栗潇远 WANG Zhengxu;LI Xiaoyuan

机构地区：[1]浙江大学公共管理学院 [2]同济大学政治与国际关系学院

出　　处：《社会科学》2023年第8期17-28,共12页Journal of Social Sciences

基　　金：上海市浦江人才计划(编号:2020-000007)资助。

摘　　要：实证社会科学研究目前存在两个主要问题:过分强调量化方法的优越性;产生了大量低水平的质性研究。造成这些问题的原因是一些研究者对两类方法的方法论基础,以及两类方法对因果推断的不同效用仍缺乏足够的理解。量化研究对数据的质量和分析的方法有很高的要求,这是许多学者没有意识到的;现有量化研究在因果效应识别上存在诸多缺陷和挑战,需要加以改进,方向是清楚的。与量化研究相比,质性研究,尤其是比较案例研究和过程研究在确认因果关系和因果机制上具有突出的优势,但需要研究者依据研究目标进行科学的案例选择和研究设计。方法论和研究设计的前沿,是围绕因果推断的研究任务,探索统一因果关系理论及其可能的实现路径,例如多元主义方法论、融合两种实验法的逻辑路径,以及以嵌套分析为代表的技术路径等。Empirical research in social sciences often overlooks a number of methodological pifalls and a lack of rigorous understanding of the methodological foundations and requirements.The main reason for this is that some researchers lack understanding of the methodological basis of the two types of methods and their different utility at the level of causal inference.Quantitative research has high requirements on data quality and analysis methods,and there are many challenges in the identification of causal effects,which need to be accounted for.Compared with quantitative research,qualitative research has special advantages in confirming causal relationship,exploring causal mechanism and realizing causal inference through structural comparison,case selection and process tracking.However,researchers need to select cases appropriately and employ rigourous research design based on different research objectives.Aiming at causal inference,the academic community is also exploring the methodological path toward a unified causality theory,such as mixed methodology,integration of the logics of two experimental methods,and nested analysis.

关 键 词：实证社会科学 因果推断 因果识别风险 统一因果关系理论 

分 类 号：C3[社会学]