
Message/Author 

Marcus Fila posted on Tuesday, January 24, 2012  6:15 am



Hi, I've been using MPLUS for a couple of months now. Today I ran a very standard kind of SEM using continuous DVs and several latent variables. However, the output did not give me all the usual model fit data (e.g., ChiSq, CFI etc.) I only got this: TESTS OF MODEL FIT Loglikelihood H0 Value 16588.837 H0 Scaling Correction Factor 1.327 for MLR Information Criteria Number of Free Parameters 111 Akaike (AIC) 33399.675 Bayesian (BIC) 33825.014 SampleSize Adjusted BIC 33472.899 (n* = (n + 2) / 24) Does anyone know how to: 1) Interpret the above, and 2) make MPLUS spit out the other model fit data? Many thanks, Marcus 


When means, variances, and covariances are not sufficient statistics for model estimation, chisquare and related fit statistics are not available. This can happen with a factor model with categorical factor indicators. In this case, nested models are often compared using loglikelihood difference testing where 2 times the loglikelihood difference is distributed as chisquare. 

Marcus Fila posted on Thursday, January 26, 2012  5:34 am



Thanks Linda. My factor indicators, however, are all quantitative. Is there thus a problem with my model? If not, would reporting the fit data I 'did' get in my output (see original post) be sufficient? Many thanks Marcus PS  I posted this again by accident. Please ignore this other post. 


Please send the output and your license number to support@statmodel.com so I can see the full situation. 

Marcus Fila posted on Thursday, January 26, 2012  8:53 am



Thanks. I just sent it. Best, Marcus 

Back to top 

