TLI和IFI大于1怎么解释？

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

X2	DF	X2/ df	RMSEA	GFI	AGFI	TLI	NFI	CFI	IFI
0.068	1	0.068	0.000	1.000	0.996	1.151	0.999	1.000	1.013

我的论文做三者之间关系模型，但是得出来TLI、IFI的值大于1，书上写是不能大于1的，但是吴明隆的书上有个例子模型和我的一样，也是大于1，上面写是可以接受的，请问这个应该怎么解释啊？

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏1 回帖

关键词：IFI 大于1 TLI RMSEA RMSE 论文 模型

As I see from your diagram this is path analysis and not a CFA model because I can't see latent variable indicators. So, you don not need to see any of the modifications indices. This is the reason, in my opinion, where you have seen these strange results. Moreover IFI can be bigger than one but in very special conditions. About TLI I have nowhere read that can be bigger than one. You may have in mind that

a. TLI is sensitive to model missipecifications;
b. IFI is not recommended for routine use.

mssr 发表于 2013-4-13 20:24
As I see from your diagram this is path analysis and not a CFA model because I can't see latent vari ...

这个不是验证性因素分析，是路径分析啊

mssr 发表于 2013-4-13 20:24
As I see from your diagram this is path analysis and not a CFA model because I can't see latent vari ...

And in what special cionditions can IFI bigger than one?Can you explain that to me?Thank you very much.

Look at this article in the following link.

http://www.uri.edu/research/cprc ... ctural%20Models.pdf

See 240 page the 1st paragraph ".... can fall ouside the 0-1 range".

Moreover look at 243 page Sample Size=200 (IFI=1.001)

这可能是由于数据四舍五入的舍入误差造成的。

你好，我遇到了和你一样的问题，请问你后来怎么解决的？吴明隆书上的例子具体在哪一页？谢谢。

我也遇到了类似的问题，应该是可以的

吴明隆的书上(P49)有说TLI,CFI,IFI值可能大于1

所以楼主最后怎么解决的？