Message/Author |
|
Anonymous posted on Tuesday, August 07, 2012 - 1:14 am
|
|
|
I have run some mixture models with constraints. They converge perfectly, although the Modification Indices are 999 - hence cannot be determined. I was wondering what are some of the possible reasons why in these instances the MI cannot be calculated? |
|
|
A 999 for a modification index means that the denominator approaches zero and the MI is not estimable. This does not reflect poorly on your model. It is known that MI's do not behave well for mixture models. |
|
Anonymous posted on Wednesday, August 08, 2012 - 1:10 am
|
|
|
Thank you. Do you have any references regarding the behaviour of MI in mixture models? |
|
|
No, this statement comes from our observations. |
|
Back to top |