Over vs just identified models PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 John posted on Wednesday, May 04, 2016 - 1:07 pm
How can I make a just-identified model where I have 1 continuous IV, 5 continuous mediators, and 1 continuous DV an over-identified model that will give me useful fit indices?
 Bengt O. Muthen posted on Wednesday, May 04, 2016 - 6:33 pm
You don't want to make it over-identified just to get fit indices. If your theory says that all paths should be in the model, then accept that you don't get model fit.
 Samantha-Kaye Johnston posted on Wednesday, June 20, 2018 - 3:00 am
Hi Bengt,

I just want to clarify your suggestion (above) regarding just-identified models.

If the just-identified model seems theoretically valid, would it be appropriate to interpret the results without the fit indices?

If yes, is there a common way to write this in the results section. (e.g., although the model is saturated, the results can still be interpreted)

Thank you

Samantha
 Bengt O. Muthen posted on Wednesday, June 20, 2018 - 11:56 am
Q1: Yes.

Q2: Most mediation models are just-identified. See e.g. the mediation book by Hayes. It is typically not commented on but you can certainly note in passing that the model is just-identified so that no overall model fit index is available.
 Samantha-Kaye Johnston posted on Wednesday, June 20, 2018 - 12:35 pm
Thanks Bengt.
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: