Differences MLR MLF PreviousNext
Mplus Discussion > Structural Equation Modeling >
 liesbeth mercken posted on Monday, August 28, 2006 - 4:05 am
I analysed my data using the MLR-method, but this method does not give me CLI, TLI, RMSEA (also not when using H1 option). I therefore analysed the same data/same model with the MLF method to obtain CLI, TLI and RMSEA. Is this a reliable indication of model fit, knowing that the best method was MLR due to the structure of my data (missing values and non-normality)?
Or is there a better way to report on model fit without using two different methods?

thank you,
kind regards,
 Linda K. Muthen posted on Tuesday, August 29, 2006 - 9:02 am
I am surprised that MLF gives chi-square when MLR does not. I would use MLR. When chi-square and related test statistics are not available, the testing of nested models using -2 times the difference in the loglikelihoods is generally used. This value is distributed as chi-square.
 liesbeth mercken posted on Wednesday, July 11, 2007 - 3:29 am
my question is related to the question I asked about a year ago. Sometimes you do not want to compare nested models, but just want to give an indication that your proposed model fits well. Is there any way to comment on model fit when MLR does not report CLI, TLI, RMSEA?

Thank you,
kind regards,
 Linda K. Muthen posted on Wednesday, July 11, 2007 - 6:14 am
For continuous outcomes, you can look at the residuals given in TECH7. For categorical outcomes, you can look at the residuals given in TECH10.
Back to top
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Options: Enable HTML code in message
Automatically activate URLs in message