|
|
FIt Indices for categorical outcomes ... |
|
Message/Author |
|
Anonymous posted on Friday, May 14, 2004 - 8:58 am
|
|
|
I do not yet have Mplus 3 so these questions may not be reasonable. a. Is a probit link-function allowed with FIML? b. What type of fit indices does Mplus3 compute with categorical outcomes with FIML? c. How is RMSEA/WRMR computed with FIML? |
|
|
No, we do not have a probit link with maximum likelihood. With maximum likelihood and categorical outcomes, you obtain the loglikelihood, AIC, BIC, sample-size adjusted BIC, Pearson chi-square, and the likelihood ratio chi-square. RMSEA and WRMR are not computed for categorical outcomes and maximum likelihood. For continuous outcomes, you can find these formulas in the technical appendices which are found at www.statmodel.com. |
|
Back to top |
|
|