after computing several nonrecursive models during the last week, I am stuck with two questions. I hope you can help me finding the corresponding answers:
Firstly, I wonder if there is a possibility to compute the stability index with MPlus, i.e. is there any possibilty to save the corresponding matrix in the output?
Kaplan/Harik/Hotchkiss (2001:335) summarize the computation of two alternative stability indices: "The stability index [...] is the largest eigenvalue of the B*B*' matrix (Equation 15.10), where B* is a matrix of diagonal elements which are set to zero. The second reported stability index is calculated as the largest eigenvalue of the same matrix B* (Equation 15.9), as suggested by Bentler & Freeman (1983)."
Secondly - after trying to reproduce the total effects in the nonrecursive case -, I wonder, if there are any references discussing that the standardized total effects might take values above 1, because of the loop-enhancing effect? Do you have ever experienced that?
Mplus does not compute the stability index. I think B* is I minus B where B is the beta matrix. You can save the model results including the beta matrix using the RESULTS option of the SAVEDATA command. TECH1 tells you which parameters are the beta matrix.
Regarding your second question, you can check the Bollen book. If you don't find an answer there, I would suggest posting the question on SEMNET.