Chi-square value in large data set PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 victor china posted on Tuesday, June 19, 2007 - 7:46 am
i preformed an analysis on a large data set (~2000 rows) with both binary and continuose variables.

Chi-Square Test of Model Fit

Value 12.161*
Degrees of Freedom 5**
P-Value 0.0326

while the rest of the indices are:
CFI/TLI

CFI 0.997
TLI 0.991

Number of Free Parameters 16

RMSEA (Root Mean Square Error Of Approximation)

Estimate 0.028

WRMR (Weighted Root Mean Square Residual)

Value 0.619

can i consider the model as significant given the CFI/TLI/RMSEA values or must i have all indices significant?
 Linda K. Muthen posted on Tuesday, June 19, 2007 - 9:18 am
These indices would seem to indicate adequate model fit. You could do a sensitivity analysis where you add a few parameters suggested by modification indices and see if doing this changes the parameter estimates for the original model.
 victor china posted on Tuesday, June 19, 2007 - 11:33 am
this is the best fit i got (after using the modifications suggested by the software, and all parameters have an effect size greater then 2.. thats the end of the road for this specific model, and the interpretation is preaty logical, but i dont think that there is a way to improve the fit (unless it is some kind of a local maxima).

thanks for the quick replay.
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: