Mulitple Group Categorical Fit Statis... PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 Kristin Ojala posted on Wednesday, March 11, 2009 - 4:37 pm
Hi,

I'm running a multiple group SEM model with two latent variables and one observed variable. The latent factors have binary outcomes. When I fit the model, the chi-square p-value is zero and the CFI=1.00, TLI= 1.03 and RMSEA=0.000. I'm not sure how to interpret this. Should I be concerned that they are so high?


Variable:
Names = DWanted Age Negative WholeEmo Romantic HusBf Knew PAge;
Usevar = Age Negative WholeEmo Romantic HusBf Knew older;
Missing = all(9999);
Categorical = Romantic HusBf Knew older;

Define:
if page==3 then older=1;
if page==2 then older=0;
if page==1 then older=0;
wholeemo=wholeemo/2;

Model: Emotion by Negative WholeEmo Romantic;
Partner by HusBf Knew older;
Emotion on Partner Age;
Age on Partner;

Output: Sampstat Standardized Residual modindices;
 Linda K. Muthen posted on Wednesday, March 11, 2009 - 5:17 pm
Your model must have no degrees of freedom. In this case, model fit is not relevant.
 Kristin Ojala posted on Friday, March 27, 2009 - 7:12 pm
Hi Linda,

I donít believe the model is just identified. According to the output I have it says I have 23 free parameters and my Chi-square has 6 degrees of freedom. The first model I ran (ignored groups) and the fit statistics seemed fine: Chi-square (3) = 8.557, CFI = .914, RMSEA = .1. I then fit the same model to each group and two of the groups separately and reported fit statistics of CFI = 1 and RMSEA = 0, whereas the third group fit statistics indicated poor fit (e.g., not CFI=1).

Would there be something that commonly causes the CFI to equal 1 and the RMSEA to equal 0 when factor indicators are categorical? When I run any of the models without a categorical statement, I donít have this problem.

Thanks.
 Linda K. Muthen posted on Saturday, March 28, 2009 - 10:13 am
I have not seen this unless the model is just-identified. Please send the output and your license number to support@statmodel.com.
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: