结构方程模型lecture3 SEM的设定和 与识别.pptVIP

SEM的设定与识别 矩阵表达 协方差阵 识别 模型设定 At the heart of the endeavor are the algebraic models. SEM的矩阵表达 SEM的协方差 SEM的设定 SEM的识别 本讲内容 SEM的矩阵表达 LISREL模型（J?reskog） 结构模型+测量模型 八大矩阵 LISREL模型的特点 An important aspect of the LISREL approach is that, in using it, variables must be arranged according to type, variables are used in different places in different equations. Manifest and latent exogenous and endogenous 问题 The LISREL equations do not explicitly include direct representation of a path in which an arrow goes from a manifest exogenous variable to a latent endogenous variable. To set up a LISREL model using the basic system, you have to input all the model matrices using a simple command language. The key is, you have to dissect the path diagram and determine which variable is an η, which is a ξ, and so on. LISREL as a system was restrictive, overly complicated, and not particularly insight provoking. SEM的矩阵表达（续1） Covariance Structure Analysis model (McDonald, 1978) 协方差结构多数能够表示成矩阵的乘积，其列元素是元素的权重或因子载荷 可以是逆矩阵的形式 Pk=Hk-1 可以是逆矩阵的形式Fki=Hki-1或Fki= (I-Hki) -1 矩阵元素可以是参数和常数 一个参数可以是其他参数的函数 Most commonly known models for analyzing covariance structures can be presented as special cases of COSAN, such as principal components analysis, confirmatory and exploratory factor analysis, and LISREL. SEM的矩阵表达（续2） Reticular Action Model （McArdle, 1978） 只需区分显变量和潜变量 显变量 潜变量 系数矩阵 fij: vj→vi 对角元素为零 RAM的协方差阵 记P=E(rr’)，令W=E(vv’)，Σ=E(mm’)，则RAM中显变量的协方差结构为： Filter matrix COSAN RAM的特点 优点 It allowed path models to be grasped and fully specified in their simplest form — as linear equations among manifest and latent variables. Instead of 8 model matrices, and a plethora of different variable types, one only needed 3 matrices. 缺点 The RAM model is somewhat wasteful in terms of the size of some of its matrices. 其他表示方法 区分随机扰动与外生变量 将随机扰动视为外生变量（James, Mulaik, and Brett, 1982） JMB模型的协方差阵 SEM的设定 参数类型 固定参数（fixed parameter）：根据假设，取事先设定数值，其数值在参数估计过程中不改变 自由参数（free parameter）：数值在参数估计的迭代过程中自由改变，直至拟合函数在固定参数的条件下达到最优 限制参数（yo