摘要
一些抽象概念,如信任、动机、满意度,很难直接测量,而需采用多个指标来进行间接测量。传统的统计方法如回归分析难以有效测量多指标、多变量之间的关系,结构方程模型(SEM)的应用则能解决这个问题。结构方程模型包括两个部分:测量方程与结构方程。结构方程模型具有多个优点,如同时处理多个因变量、允许自变量与因变量均含有测量误差、衡量整个模型的拟合优度等。最后给出了一个采用结构方程模型来分析网上信任问题的实例。
A few abstract concepts, such as trust, motivation and satisfaction, are difficult to be directly measured and can only be measured with multiple indicators. Traditional statistics methods, such as regression analysis, cannot effectively measure relationships between multiple indicators and multiple variables. However, structural equation modeling （SEM） can be used to solve the problem. SEM includes two parts： measurement equation and structural equation. SEM has many advantages, such as simultaneously dealing with multiple dependent variables, allowing both independent and dependent variables having measurement errors, and measuring the fitness of the whole model. The study finally provides an example for applying SEM to analyze Online Trust.
出处
《工业工程与管理》
2006年第5期99-102,共4页
Industrial Engineering and Management
基金
国家社科基金项目(06BJY101)
关键词
结构方程模型
网上信任
Structural Equation Modeling
online trust
作者简介
周涛（1979-），男，湖北荆门人，博士研究生，主要研究方向为电子商务与电子市场。
