luhua zhao posted on Tuesday, June 20, 2006 - 7:27 pm
I am thinking of getting mplus, but currently still using sas for factor analysis. We have 31 variables, all are 3 level categorical variables. Based on sas techinical support, I get the polychoric correlation matrix, and used this matrix in proc factor.
I wonder if this approach is the same one used in mplus.
Besides, what are the major advantages of using mplus? Thanks.
Mplus uses this approach with the unweighted least squares estimator. With the weighted least squares estimator, a weight matrix is involved. With both estimators, standard errors and chi-square fit and other fit statistics are available.