Chi-Square Difference Testing Using t... PreviousNext
Mplus Discussion > Confirmatory Factor Analysis >
 Len Burns posted on Tuesday, May 21, 2002 - 5:21 pm
Thank-you for your earlier response to my question. I am currently comparing some output from MPLUS with EQS version 6.0 in regard to testing the difference between nested models with the Satorra-Bentler Scaled Chi Square.

I have discovered a difference in output between the programs--the MLM estimation procedure in MPLUS and the Robust Estimation procedure in EQS. The difference concerns the Satorra-Bentler Scaled Chi-Square. Below is the relevant output (this is a CFA analysis--two factors, eight items per factor, the purpose being to see if the two factor model is better than the one factor model)

Model 1 Model 2 (2 factor)
ML Estimation 4613.46 1493.975
MLM Estimation 2244.893 740.198
df 104 103
SCF 2.055 2.018
cd = 5.866
TRd= 531.79

Model 1 Model 2
ML Estimation 4611.328 1493.318
Robust Estimation 1924.870 703.471
df 104 103
SCF 2.396 2.123
cd = 30.52
TRd = 102

All the other output is the same across the two programs except for the Satorra-Bentler Scaled Chi-Square. My guess is that the MLM estimation procedure in MPLUS is not exactly the same as the Robust estimation procedure in EQS. However, I am uncertain about this. In addition, the values that I calculate for the TRd make "sense" with the MPLUS output. The values for TRd, however, do not make "sense" for the EQS output.

I would greatly appreciate any thoughts you might have on this. Thanks so much.

Len Burns
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