We met at the M-Plus course at Hopkins in which I showed you my output and you told me I could e-mail you. I am running a model in which I am looking whether first grade agression, inpulsivity and hyperactivity (measured in the Fall and Spring of first grade) predict 11th grade gambling, alcohol use, conduct disorder and illegal drug use (all categorical outcomes). As I told you, Katherine Masyn is helping me with the modeling. The first grade model is only identified when we constrain factor loadings to be equal whithin factors and when we allow for covariance between same measures. The problem occurs when I try to regress the distal outcomes on the 3 factors derived from 1st grade data. As you suggested, I increased the Miterations in the model and I also ran separate models with each of the distal outcomes at a time and with each of the factors at a time. When I run these models, all of them rubn well with the exception of the model with only f3 in which i get an error message: THE ESTIMATED COVARIANCE MATRIX COULD NOT BE INVERTED. COMPUTATION COULD NOT BE COMPLETED IN ITERATION 1472. CHANGE YOUR MODEL AND/OR STARTING VALUES. Also, I get a similar message when I try to run the final model (with the 3 factors and the 4 categorical distal outcomes),when I increase the miterations up to 3000. What do you suggest I do? How do I send you the output?