【作者】 褚荣伟；

【Author】 Rongwei Chu (Department of Business Administration, School of Management, Fudan University, Shanghai 200433)

【机构】 复旦大学管理学院企业管理系；

【摘要】 利用数据来发现社会科学的规律已经成为研究的一种主流,变量之间的关系也越来越受到学者们的关注,而社会科学具有天然的情境依赖性,因此这样的关系必然具有不同的机制和实现环境。本文希望区分调节作用和中介作用,并分别提出基本的统计检验方法。调节变量是指影响两个变量之间关系的第三个变量,而中介变量则是变量之间的关系机制,也就是中介变量本身就是变量之间的关系体现。调节作用可以通过调节回归分析和结构方程模型来得以检验,中介作用则可以通过一组回归方程的系数显著性和Sobel检验方法来检验。文章接着从应用的角度,简单回顾了市场营销研究中在过去二十几年的研究中对于此统计方法的重视和使用程度,结果发现已经有越来越多的营销学者为了更深入的探讨营销现象,开始将这样的概念引入其研究之中,但是市场营销方面的研究基本上还处于简单应用的阶段,对于方法论上面的贡献是不足的,而且也缺乏相对应的回顾和评价。更多还原

【Abstract】 That more data are involved has become a kind of research mainstream, especially when digging out the underlying rule of social sciences. All researchers are putting more emphases on the relationship among the kinds of variables. What’s more, social science is inherently context-specific, which results that such relationships will vary with different conditions and obey with different mechanism. This paper aimed to differentiate moderators and mediators and propose the testing methods respectively. Moderator was the third variable which influenced the relationship between independent variables and dependent variables in terms of magnitude and direction. Mediators disclosed the mechanism between independent variables and dependent variables. We could use moderated regression analysis and structural equation model to test the moderating effect. On the other hand, a series of regression equations and Sobel statistics were adopted to confirm the existing of mediators. After that, this paper depicted the status in quo of application in marketing research through the basic literature review. The results show that though moderating and mediating effect have aroused the attention for more researchers, they just apply them with little contribution to methodological consideration.更多还原