结构方程模型应用中样本和参数估计问题探析

Analysis on the Samples and Parameter Estimations in the Application of Structural Equation Modeling

结构方程模型整合多种传统统计方法的优点,越来越多地被运用到社会、心理、管理、医学、教育等学科领域。然而该方法对样本选取的要求高,参数估计方法灵活。如果运用者只是机械地生搬硬套,很容易导致误用,从而产生错误的分析结果。本文对结构方程模型分析中的样本数量和参数估计方法进行探讨,以有效提高该方法的运用效率,提高研究者的实证研究水平。

Structural equation modeling integrates the advantages of various traditional statistical methods, and it is more and more applied to many fields of society, such as psychology, manage- ment, medicine, education and so on. However, this method is very strict to the sample selection criteria, and the methods of pa- rameter estimation are also very flexible. If investigators use the method only mechanically, it is easy to lead to misuse, which leads to erroneous analysis results. In this paper, we discuss the samples and parameter estimation methods in the structural e- quation modeling, in order to effectively improve the efficiency of the method, and to improve the level of empirical research.