结构方程论坛SEM-Structural·Equation·Modeling's Archiver

dataanlysis 发表于 2011-7-2 23:50

single arrow stands for causal relation and double arrow stands for correlation.<br /> normally, you would run a preliminary model with all the latent variables correlated.<br /> if the model is identified, that means your causal model will be identifed, too.<br /> as to the causal relationship, that is, the direction of the arrows need to be guided by theories and previous studies.<br /> you can also use modification index as help to add or remove the paths, but still the model needs to be according to theories instead of SEM program.

偏最小二乘 发表于 2011-7-2 23:55

<strong> 21# <i>qoop</i> </strong> <br /> <br /> 谢谢~~~~~~~~~~~~~~~~~~~~~

华南心理学 发表于 2011-7-3 00:00

在做此类的FULL SEM之前,检验用来测量各个latent variable的问卷或工具的measurement model是一个很重要的步骤,这对整个struatural model的模型拟合和结果解释具有重要的影响。如果楼主没有做过这一步骤,可以尝试一下。

潜变量分析 发表于 2011-7-3 00:05

<strong> 23# <i>liuliqingdong</i> </strong> <br /> <br /> 谢谢

51jijin 发表于 2011-7-3 00:10

感觉你变量间的关系太多了,由于待估协方差个数太多,样本容量如果不够,肯定拟合效果不好。

因子分析 发表于 2011-7-3 00:15

<strong> 25# <i>mzl79</i> </strong> <br /> 谢谢

spsschina 发表于 2011-7-3 00:20

学习一下 <img src="http://bbs.pinggu.org/images/smilies/grapeman/08.gif" smilieid="48" border="0" alt="" />

因子分析 发表于 2011-7-3 00:25

觉得你的外生潜变量相关关系没有画全,<br /> 既然外生潜变量邻接的两个两两相关,必然全部两两相关,把所有相关关系限定加上去,拟合质量会更高的。<br /> 有相关关系的外生潜变量,可能说到某些共同因子的作用,不限定这些相关性,模型拟合的适度会急剧下降。

潜变量分析 发表于 2011-7-3 00:30

看不懂。。今晚上开始学lisrel,我果真是学前班的 <img src="http://bbs.pinggu.org/images/smilies/default/em06.gif" smilieid="140" border="0" alt="" />

页: 1 [2]

Powered by Discuz! Archiver 7.2  © 2001-2009 Comsenz Inc.