sunny shi posted on Thursday, May 14, 2009 - 8:16 am
Dear Mplus, When I was running a linear growth model with a correlated within-person residual covariance structure, I got the following error message:
THE MODEL ESTIMATION TERMINATED NORMALLY
THE STANDARD ERRORS OF THE MODEL PARAMETER ESTIMATES MAY NOT BE TRUSTWORTHY FOR SOME PARAMETERS DUE TO A NON-POSITIVE DEFINITE FIRST-ORDER DERIVATIVE PRODUCT MATRIX. THIS MAY BE DUE TO THE STARTING VALUES BUT MAY ALSO BE AN INDICATION OF MODEL NONIDENTIFICATION. THE CONDITION NUMBER IS 0.990D-12. PROBLEM INVOLVING PARAMETER 14.
As the default method is ML, the information matrix is observed. However, I am wondering what causes the untrustworthy standard errors? What the untrustworthy comes from? Should I delete all these untrustworthy results? ( I am running 1000 replications)