结构方程模型是心理统计学中用来探讨多变量之间关系的一种重要统计方法。结构方程模型的优势表现为:能够判断模型拟合程度的测量、预先假定变量间因果关系、包含潜在变量、处理测量误差、进行模型整体评价、探讨系统内多个变量间关系、模型构建等。为了符合心理学研究抽样调查较多的实际情况,并且在小样本情况下能得到更稳健的结果,结构方程模型的贝叶斯方法使用先验信息,能得到更准确的参数估计、潜在变量估计以及用于模型比较的统计量。在心理学研究的应用过程中应注意结构方程模型的前提假设、自由度和拟合指数等问题。
Structural equation model was used to investigate the psychological relationship between significant multivariate statistical methods .The advantages of structural equation model are as follows:determining the extent of the measurement model fit , presupposing a causal relationship between variables , including the latent variables , handling measurement error , overall evaluating the model , exploring the relationships between multiple variables within the system , constructing model , and so on .To meet the practical situation of the psychological research sam-pling , and to achieve more robust results in small samples , Bayesian methods of structural equation modeling uses a priori information to get a more accurate parameter estimates , latent variable estimates and statistics for the model comparison .In the application process of psychological research , attention should be paid to such aspects as the premise of structural equation modeling assumptions , freedom and fit indices .