Anonymous posted on Tuesday, October 05, 2004 - 9:21 am
I'm doing a CFA with MPLUS v 3.0 and I have defined 2 item parcels for each of 7 factors. Each parcel is the mean of 2-3 binary items. I do not get a solution. Either I get he message that there is not a positive definite residual covariance matrix or I if I let all the factor loadings be free and fix the factor variances to 1, the model does not converge. Any suggestions?
Also, how can I set the starting values to be the end values of the first analysis?