

Fit indices for CFA with categorical ... 

Message/Author 


I am aware that a CFA with categorcal indicators, using ULS estimation, only provides the chisquared statistic as a measure of goodness of fit. Is there any way to calculate additional fit indices? 


I just ran a CFA with categorical indicators using ULS and I get all fit statistics. Perhaps you are using an older version of Mplus. 


Hi! I ran a CFA model with four factors and categorical (ordered) indicators. The fit statistics that I got are shown below (copied from the output). It says that THE CHISQUARE TEST IS NOT COMPUTED BECAUSE THE FREQUENCY TABLE FOR THE LATENT CLASS INDICATOR MODEL PART IS TOO LARGE. I wonder about how I can determine the fit of my model to the data (just like I do with CFI, RMSEA, or Chisquare) since I do not compare this model to any other models here. Thank you!! Number of Free Parameters= 62 Loglikelihood H0 Value 16281.545 Information Criteria Akaike (AIC) 32687.090 Bayesian (BIC) 33009.470 SampleSize Adjusted BIC 32812.523 (n* = (n + 2) / 24) 


It sounds like you are are using maximum likelihood estimation. In this case, chisquare and related fit statistics like RMSEA, CFI, etc. are not available. Two chisquares comparing observed versus estimated frequencies are computed when the frequency table is not too large. Generally these values should not be interpreted for more than 8 categorical indicators and should not be interpreted if the two do not agree. There are no absolute fit statistics in this situation. Nested models can be compared by using 2 times the loglikelihood difference which is distributed as chisquare. 

Back to top 

